We can't find the internet
Attempting to reconnect
Something went wrong!
Hang in there while we get back on track
Simulation of Chlorophyll a Concentration in Donghu Lake Assisted by Environmental Factors Based on Optimized SVM and Data Assimilation
Summary
An optimized machine learning model was developed and combined with data assimilation techniques to simulate chlorophyll-a concentrations — an indicator of algal growth — in Donghu Lake, China. The model accurately reproduced observed chlorophyll patterns by incorporating environmental factors like temperature and nutrients. Better lake eutrophication models support water quality management and early warning of harmful algal blooms.
Lake eutrophication is a global water environmental problem and has become a research focus nowadays. Chlorophyll a concentration is an important index in terms of evaluating lake eutrophication. The aim of this study was to build an effective and universal empirical model for simulation of chlorophyll a concentration in Donghu Lake. On the basis of the relationship between chlorophyll a concentration and dissolved oxygen (DO), water temperature (T), total nitrogen (TN), and total phosphorus (TP), models for simulating chlorophyll a concentration were built by using simulated annealing (SA), genetic algorithm (GA), artificial bee colony (ABC), and particle swarm optimization (PSO) to optimize parameters of support vector machine (SVM). Moreover, a collaborative mode (Col-SVM model) was built by introducing data assimilation, and meanwhile, accuracy and universality of the model were studied. Modeling results showed that the application of optimization algorithms and data assimilation improved the performance of modeling based on SVM. Model simulation results demonstrated that the Col-SVM model has high accuracy, decent stability, and good simulation effect; the root mean square error (RMSE), mean absolute percentage error (MAPE), Nash–Sutcliffe efficiency coefficient (NSE), bias, and mean relative error (MRE) between simulated values and observed values were 10.07 μg/L, 0.31, 0.96, −0.050, and 0.15, respectively. In addition, model universality analysis results revealed that the Col-SVM model has good universality and can be used to simulate the chlorophyll a concentration of Donghu Lake at different times. Overall, we have built an effective and universal simulation model of chlorophyll a concentration that provides a new idea and method for chlorophyll a concentration modeling.
Sign in to start a discussion.
More Papers Like This
Chlorophyll-a Detection Algorithms at Different Depths Using In Situ, Meteorological, and Remote Sensing Data in a Chilean Lake
Researchers used a combination of field measurements, weather data, and satellite imagery to estimate chlorophyll-a concentrations at different depths in a Chilean lake. They compared deep learning and statistical models and found all three approaches performed well for predicting algal levels in the freshwater ecosystem. The study advances water quality monitoring techniques that can help track environmental changes, including those potentially linked to pollution.
Chlorophyll-a unveiled: unlocking reservoir insights through remote sensing in a subtropical reservoir
Researchers used satellite data from Landsat-8 and Sentinel-2 combined with machine learning to estimate chlorophyll-a concentrations — a measure of algae levels — in a South African reservoir. This remote sensing approach enables water managers to monitor reservoir health continuously without costly field sampling, helping detect harmful algal blooms earlier.
Regional Satellite Algorithms to Estimate Chlorophyll-a and Total Suspended Matter Concentrations in Vembanad Lake
Researchers developed regional satellite algorithms to estimate chlorophyll-a concentrations and total suspended matter in Vembanad Lake, India, using remote sensing data to monitor water quality in a highly productive but increasingly polluted coastal ecosystem. The algorithms were calibrated against in-situ measurements and found to improve the accuracy of water quality assessments compared to global ocean-color models, supporting sustainable development monitoring goals.
Water Quality Grade Identification for Lakes in Middle Reaches of Yangtze River Using Landsat-8 Data with Deep Neural Networks (DNN) Model
Researchers developed a deep neural network model applied to Landsat-8 satellite data to automatically identify water quality grades for lakes in the middle Yangtze River reaches, demonstrating that machine learning and remote sensing can provide cost-effective large-scale monitoring as an alternative to labor-intensive in situ measurements.
Response Surfaces Method and Artificial Intelligence Approaches for Modeling the Effects of Environmental Factors on Chlorophyll a in Isochrysis galbana
Researchers optimized chlorophyll a production in the microalga Isochrysis galbana using Box-Behnken Design, finding maximum content of 0.51 mg/L at salinity 30 ppt, pH 8.0, and 72.1 mg/L nitrogen, then compared response surface methodology with three AI models — ANN, SVM, and RBFNN — for predicting chlorophyll a content. The AI models achieved R2 values of 0.91–0.92, outperforming the RSM model in prediction accuracy.