An Intelligent Early Warning System for Harmful Algal Blooms: Harnessing the Power of Big Data and Deep Learning.

Harmful algal blooms (HABs) pose a significant ecological threat and economic detriment to freshwater environments. In order to develop an intelligent early warning system for HABs, big data and deep learning models were harnessed in this study. Data collection was achieved utilizing the vertical aquatic monitoring system (VAMS). Subsequently, the analysis and stratification of the vertical aquatic layer were conducted employing the "DeepDPM-Spectral Clustering" method. This approach drastically reduced the number of predictive models and enhanced the adaptability of the system. The Bloomformer-2 model was developed to conduct both single-step and multistep predictions of Chl-a, integrating the " Alert Level Framework" issued by the World Health Organization to accomplish early warning for HABs. The case study conducted in Taihu Lake revealed that during the winter of 2018, the water column could be partitioned into four clusters (Groups W1-W4), while in the summer of 2019, the water column could be partitioned into five clusters (Groups S1-S5). Moreover, in a subsequent predictive task, Bloomformer-2 exhibited superiority in performance across all clusters for both the winter of 2018 and the summer of 2019 (MAE: 0.175-0.394, MSE: 0.042-0.305, and MAPE: 0.228-2.279 for single-step prediction; MAE: 0.184-0.505, MSE: 0.101-0.378, and MAPE: 0.243-4.011 for multistep prediction). The prediction for the 3 days indicated that Group W1 was in a Level I alert state at all times. Conversely, Group S1 was mainly under an Level I alert, with seven specific time points escalating to a Level II alert. Furthermore, the end-to-end architecture of this system, coupled with the automation of its various processes, minimized human intervention, endowing it with intelligent characteristics. This research highlights the transformative potential of integrating big data and artificial intelligence in environmental management and emphasizes the importance of model interpretability in machine learning applications.

Simulating chlorophyll-a fluorescence changing rate and phycocyanin fluorescence by using a multi-sensor system in Lake Taihu, China.

Algal pollution in water sources has posed a serious problem. Estimating algal concentration in advance saves time for drinking water plants to take measures and helps us to understand causal chains of algal dynamics. This paper explores the possibility of building a short-term algal early warning model with online monitoring systems. In this study, we collected high-frequency data for water quality and weather conditions in shallow and eutrophic Lake Taihu by an in situ multi-sensor system (BIOLIFT) combined with a weather station. Extracted chlorophyll-a from water samples and chlorophyll-a fluorescence differentiated according to different algal classeses verified that chlorophyll-a fluorescence continuously measured by BIOLIFT only represent chlorophyll-a of green algae and diatoms. Stepwise linear regression was used to simulate the chlorophyll-a fluorescence changing rate of green algae and diatoms together (ΔChla-f%) and phycocyanin fluorescence concentration (blue-green algae) on the water surface layer (CyanoS). The results show that nutrients (total N, NO3-N, NH4-N, total P) were not necessary parameters for short-term algal models. ΔChla-f % is greatly influenced by the seasons, so seasonal partition of data before modeling is highly recommended. CyanoSmax and ΔChla-f% were simulated by only using multi-sensor and meteorological data (R2 = 0.73; 0.75). All the independent variables (wave, water temperature, relative humidity, depth, cloud cover) used in the model were measured online and predictable. Wave height is the most important independent variable in the shallow lake. This paper offers a new approach to simulate and predict the algal dynamics, which also can be applied in other surface water.

Identifying spatio-temporal dynamics of trace metals in shallow eutrophic lakes on the basis of a case study in Lake Taihu, China.

In shallow eutrophic lakes, metal remobilization is closely related to the resuspension and eutrophication. An improved understanding of metal dynamics by biogeochemical processes is essential for effective management strategies. We measured concentrations of nine metals (Cr, Cu, Zn, Ni, Pb, Fe, Al, Mg, and Mn) in water and sediments during seven periods from 2014 to 2018 in northern Lake Taihu, to investigate the metal pollution status, spatial distributions, mineral constituents, and their interactions with P. Moreover, an automatic weather station and online multi-sensor systems were used to measure meteorological and physicochemical parameters. Combining these measurements, we analyzed the controlling factors of metal dynamics. Shallow and eutrophic northern Lake Taihu presents more serious metal pollution in sediments than the average of lakes in Jiangsu Province. We found chronic and acute toxicity levels of dissolved Pb and Zn (respectively), compared with US-EPA "National Recommended Water Quality Criteria". Suspended particles and sediment have been polluted in different degrees from uncontaminated to extremely contaminated according to German pollution grade by LAWA (Bund/Länder-Arbeitsgemeinschaft Wasser). Polluted particles might pose a risk due to high resuspension rate and intense algal activity in shallow eutrophic lakes. Suspended particles have similar mineral constituents to sediments and increased with increasing wind velocity. Al, Fe, Mg, and Mn in the sediment were rarely affected by anthropogenic pollution according to the geoaccumulation index. Among them, Mn dynamics is very likely associated with algae. Micronutrient uptake by algal will affect the migration of metals and intensifies their remobilization. Intensive pollution of most particulate metals were in the industrialized and down-wind area, where algae form mats and decompose. Moreover, algal decomposition induced low-oxygen might stimulate the release of metals from sediment. Improving the eutrophication status, dredging sediment, and salvaging cyanobacteria biomass are possible ways to remove or reduce metal contaminations.

Highly time-resolved analysis of seasonal water dynamics and algal kinetics based on in-situ multi-sensor-system monitoring data in Lake Taihu, China.

Predicting algal blooms is challenging due to rapid growth rates under suitable conditions and the complex physical, chemical, and biological processes involved. Physico-chemical parameters, monitored in this study by a high-resolution in-situ multi-sensor system and derived from lab-based water sample analyses, show the seasonal variation and have different degrees of vertical gradients across the water column. Through analyzing the changes and relations between multi-factors, we reveal pictures of water quality dynamics and algal kinetics. Nitrate has regular seasonal changes different to the seasonal patterns of total dissolved Phosphorus. Positive correlations are found between Chlorophyll a fluorescence and temperature, wind-induced resuspension and mixing promote the augment of Cyanobacteria fluorescence (Phycocyanin) signal. While the resuspension can also result in the increase of turbidity and affect the light environment for hydrophytes, the algal scums are the main reason for the high turbidity on the surface, which lower the illumination radiation in the water body. Those parameters are the primary dominants responsible for the change of algae from our monitoring data, which could be used as indicators for the dynamic changes of algae in the future.

Hyperspectral Data and Machine Learning for Estimating CDOM, Chlorophyll a, Diatoms, Green Algae and Turbidity.

Inland waters are of great importance for scientists as well as authorities since they are essential ecosystems and well known for their biodiversity. When monitoring their respective water quality, in situ measurements of water quality parameters are spatially limited, costly and time-consuming. In this paper, we propose a combination of hyperspectral data and machine learning methods to estimate and therefore to monitor different parameters for water quality. In contrast to commonly-applied techniques such as band ratios, this approach is data-driven and does not rely on any domain knowledge. We focus on CDOM, chlorophyll a and turbidity as well as the concentrations of the two algae types, diatoms and green algae. In order to investigate the potential of our proposal, we rely on measured data, which we sampled with three different sensors on the river Elbe in Germany from 24 June⁻12 July 2017. The measurement setup with two probe sensors and a hyperspectral sensor is described in detail. To estimate the five mentioned variables, we present an appropriate regression framework involving ten machine learning models and two preprocessing methods. This allows the regression performance of each model and variable to be evaluated. The best performing model for each variable results in a coefficient of determination R 2 in the range of 89.9% to 94.6%. That clearly reveals the potential of the machine learning approaches with hyperspectral data. In further investigations, we focus on the generalization of the regression framework to prepare its application to different types of inland waters.