Predicting Cyanobacterial Blooms Using Hyperspectral Images in a Regulated River.

Process-based modeling for predicting harmful cyanobacteria is affected by a variety of factors, including the initial conditions, boundary conditions (tributary inflows and atmosphere), and mechanisms related to cyanobacteria growth and death. While the initial conditions do not significantly affect long-term predictions, the initial cyanobacterial distribution in water is particularly important for short-term predictions. Point-based observation data have typically been used for cyanobacteria prediction of initial conditions. These initial conditions are determined through the linear interpolation of point-based observation data and may differ from the actual cyanobacteria distribution. This study presents an optimal method of applying hyperspectral images to establish the Environmental Fluid Dynamics Code-National Institute of Environment Research (EFDC-NIER) model initial conditions. Utilizing hyperspectral images to determine the EFDC-NIER model initial conditions involves four steps that are performed sequentially and automated in MATLAB. The EFDC-NIER model is established using three grid resolution cases for the Changnyeong-Haman weir section of the Nakdong River Basin, where Microcystis dominates during the summer (July to September). The effects of grid resolution on (1) water quality modeling and (2) initial conditions determined using cumulative distribution functions are evaluated. Additionally, the differences in Microcystis values are compared when applying initial conditions using hyperspectral images and point-based evaluation data. Hyperspectral images allow detailed initial conditions to be applied in the EFDC-NIER model based on the plane-unit cyanobacterial information observed in grids, which can reduce uncertainties in water quality (cyanobacteria) modeling.

Using convolutional neural network for predicting cyanobacteria concentrations in river water.

Machine learning modeling techniques have emerged as a potential means for predicting algal blooms. In this study, synthetic spatio-temporal water quality data for a river section were generated with a 3D water quality model and used to investigate the capability of a convolutional neural network (CNN) for predicting harmful cyanobacterial blooms. The CNN model displayed a reasonable capacity for short-term predictions of cyanobacteria (Microcystis) biomass. In the nowcasting of Microcystis, the CNN performance had a Nash-Sutcliffe Efficiency (NSE) of 0.87. An increase in the forecast lead time resulted in a decrease in the prediction accuracy, reducing the NSE from 0.87 to 0.58. As the spatial observation density increased from 20% to 100% of the input image grids, the CNN prediction NSE had improved from 0.70 to 0.84. Adding noise to the data resulted in accuracy deterioration, but even at the noise amplitude of 10%, the accuracy was acceptable for some applications, with NSE = 0.76. Visualization of the CNN results characterized its performance variations across the studied river reach. Overall, this study successfully demonstrated the capability of the CNN model for cyanobacterial bloom prediction using high temporal frequency images.

Variable update strategy to improve water quality forecast accuracy in multivariate data assimilation using the ensemble Kalman filter.

Data assimilation in complex water quality modeling is inevitably multivariate because several water quality variables interact and correlate. In ensemble Kalman filter applications, determining which variables to include and the structure of the relationships among these variables is important to achieve accurate forecast results. In this study, various analysis methods with different combinations of variables and interaction structures were evaluated under two different simulation conditions: synthetic and real. In the former, a synthetic experimental setting was formulated to ensure that issues, including incorrect model error assumption problem, spurious correlation between variables, and observational data inconsistency, would not distort the analysis results. The latter did not have such considerations. Therefore, this process could demonstrate the undistorted effects of the different analysis methods on the assimilated outputs and how these effects might diminish in real applications. Under synthetic conditions, updating a single active variable was found to improve the accuracy of the other active variables, and updating multiple active variables in a multivariate manner mutually enhanced the accuracy of the variables if proper ensemble covariance and observation data consistency were ensured. The results of the real case indicated a weakened mutual enhancement effect, and the methods in which variable localization were applied yielded the best analysis results. However, the multivariate analysis methods produced more accurate forecasting results, indicating that these methods could be superior. Therefore, it is suggested that multivariate analysis methods be considered first for water quality modeling, and the application of variable localization should be considered if significant spurious correlations and data inconsistency are present.