Estimate Water Quality Research Articles

Assessing the Efficiency of a Random Forest Regression Model for Estimating Water Quality Indicators

Assessing the Efficiency of a Random Forest Regression Model for Estimating Water Quality Indicators

Implementation of remote sensing algorithms to estimate TOC, Chl-a, and TDS in a tropical water body; Sanalona reservoir, Sinaloa, Mexico.

The present study implements a methodology to estimate water quality values using statistical tools and remote sensing techniques in a tropical water body Sanalona. Linear regression models developed by Box-Cox transformations and processed data from LANDSAT-8 imagery (bands) were used to estimate TOC, TDS, and Chl-a of the Sanalona reservoir from 2013 to 2020 at five sampling sites measured every 6 months. A band discriminant analysis was carried out to statistically fit and optimize the proposed algorithms. Coefficients of determination beyond 0.9 were obtained for these water quality parameters (r2TOC = 0.90, r2TDS = 0.95, and r2Chl-a = 0.96). A comparison between the estimated and observed water quality was carried out using different data for validation. The validation of the models showed favorable results with R2TOC = 0.8525, R2TDS = 0.8172, and R2Chl-a = 0.9256. The present study implemented, validated, and compared the results obtained by using an ordered and standardized methodology proposed for the estimation of TOC, TDS, and Chl-a values based on water quality parameters measured in the field and using satellite images.

Geostatistical appraisal of water quality, contamination, source distribution of potentially toxic elements (PTEs) in the lower stretches of Subarnarekha River (Odisha), India, and health risk assessment by Monte Carlo simulation approach.

In the present study, the status of water quality, environmental contamination in the lower stretch of Subarnarekha River with respect to potentially toxic elements (PTEs), its seasonal distribution, and ecotoxicological health impacts were investigated. For this purpose, a combination of indexing approaches and geospatial methods was used. The estimated water quality index (WQI) has shown that the river water falls under "moderate to very poor" category during the pre-monsoon and "moderate to poor" category in the post-monsoon season. The abundance of PTEs (Pb, Cu, Ni, Cd, Fe, and Cr) was on the higher side during the pre-monsoon in comparison with the post-monsoon season. The results of contamination index (Cd) and heavy metal evaluation index (HEI) explain that Subarnarekha River has low-to-moderate levels of contamination with PTEs in the majority of sampling sites. However, HPI indicated that the river water is moderate-to-highly contaminated with PTEs in both seasons. Principal component analysis (PCA) and cluster analysis (CA) reveal that anthropogenic sources are prime contributors to PTEs contamination in Subarnarekha River. The potential non-cancerous health concerns for child and adults due to Cr and Pb in some sampling stations along the river stretch have been observed. The carcinogenic risk (CR) has been established for Cr, Pb, and Cd in Subarnarekha River with Cr (> 10-4) as the most unsafe element. Monte Carlo simulation (MCS) indicates a high risk of cancer hazards due to Cr (values > 1E-04) in present as well as future for both child and adults.

Three-dimensional augmentation for hyperspectral image data of water quality: An Integrated approach using machine learning and numerical models

This research introduces a comprehensive methodology to enhance hyperspectral image data (HSD) utility, specifically focusing on the three-dimensional (3-D) augmentation of Chlorophyll-a (Chl-a). This study comprises three significant steps: (1) the augmentation of limited field water quality data in terms of time interval and number of variables using neural network models, (2) the generation of 3-D data using numerical models, and (3) the extension of the hyperspectral image data into 3-D data using machine learning models. In the first phase, Multilayer Perceptron (MLP) models were developed to train water quality interactions and successfully generated high-frequency water quality data by adjusting biased measurements and predicting detailed water quality variables. In the second phase, high-frequency data generated by MLP models were applied to develop two numerical models. These numerical models successfully generated 3-D data, thereby demonstrating the effectiveness of integrating numerical modeling with neural networks. In the final phase, ten machine learning models were trained to generate 3-D Chl-a data from HSD. Notably, the Gaussian Process Regression model exhibited superior performance, effectively estimating 3-D Chl-a data with robust accuracy, as evidenced by an R-square value of 0.99. The findings align with theories of algal bloom dynamics, further validating the effectiveness of the approach. This study demonstrated the successful integrated development for HSD extension using machine learning models, numerical models, and original HSD, highlighting the potential of such integrated methodologies in advancing water quality monitoring and estimation. Notably, the approach leverages readily accessible data, allowing for the swift generation of results and bypassing time-consuming data collection processes. This research marks a significant step towards more robust, comprehensive water quality monitoring and prediction, thereby facilitating better management of aquatic ecosystems.

Assessment and estimation of water quality using multi-linear regression

All living things require water in order to survive. Water that is drinkable for people should be safe and avoid of dangerous microorganisms. Therefore, water that has been purified should be made available to the general public before it can be utilized for human consumption. To ensure that the parameters are within allowable bounds, specific standards in the water must be investigated. Prior to use, the ground water must be evaluated and cleaned. The statistical method known as multiple linear regression makes use of two or more independent variables to forecast the value of a dependent variable. With the use of this method, we can ascertain the model’s variation as well as the proportionate contributions of each independent variable to the overall variance. The assessment of water quality can be done by conducting the pH, turbidity, conductivity, acidity, alkalinity, chloride, iron, total solids, dissolved oxygen tests. Multiple samples are collected from various locations for accuracy of the test results. The aim of the study is to assess the water quality parameters through laboratory experiments and to relate the water parameters to water quality index using the multilinear regression.

GREEN CHEMISTRY: USING MORE SUSTAINABLE TECHNIQUES TO ESTIMATE THE WATER QUALITY OF THE UBERABA RIVER USING ORBITAL OPTICAL SENSORS

Green Chemistry aims to prevent pollution resulting from activities in the chemical sector. Generally, laboratory activities generate pollution because toxic substances are used to perform analyses, which produce waste once the analyses are completed. This study aimed to monitor the quality of the Uberaba River using more sustainable and non-destructive techniques through orbital remote sensing and to develop methods for estimating water quality at reduced costs using orbital sensors and machine learning. The values of water quality parameters, such as iron, nitrate, chloride, total phosphorus, and total nitrogen, were studied between 2018 and 2023. Orbital sensing data (spectral bands and vegetation indices) were taken from the exact geographic coordinates of the collection points. Data were analyzed using Pearson’s correlation coefficient and random forest regression analysis. The study concludes that it is possible to perform orbital remote sensing by estimating water quality through random forest regression, correlating the information obtained from Sentinel-2 images with the values of these parameters. This approach is a sustainable technique that does not generate waste and represents a 100% saving compared to conventional chemical analyses.

Water quality and ecological state estimate of large lakes of Yakutsk City (Lake Saysary, Lake Sergelyakh) in the end of ice period in 2021

Water quality of water bodies situated near populated localities is an important ecological problem. To estimate water quality in Lake Saysary and Lake Sergelyakh within Yakutsk City, hydrochemical, geochemical and microbiological methods were used. By research moment, the water in these lakes did not meet quality norms established for water of water bodies for fishery use as well as for drinking and cultural-domestic water use. The data obtained make pay attention to the necessity of the control for limitation of water use from these lakes for human needs and of activities for improvement of their quality.

A Mathematical Model for Water Quality Assessment: Evidence-Based from Selected Boreholes in Federal University Dutse, Nigeria

The present study assessed the quality of water sampled from different boreholes on the campus of Federal University Dutse, Nigeria, using a mathematical modelling approach. A model for estimating water quality was developed based on physicochemical parameters such as pH, electrical conductivity, temperature, turbidity, and total hardness measured from each borehole. The correlation analysis of physicochemical data indicates a strong correlation of about 99% between the real-life data collected from six (6) different boreholes and the model’s predictions. From the results, the sensitivity analysis revealed that electrical conductivity plays the highest role in determining water quality, followed by total hardness, temperature has the third highest impact, followed by turbidity, the fourth, and pH has the least impact in determining water quality in the listed boreholes. Therefore, in any case of intervention, the water quality regulatory body should be sent regularly to the tertiary institutions in the state for routine check-ups.

Groundwater Geochemical Investigation and Quality Assessment for Drinking and Irrigation Uses in Sohagpur Coalfield, Madhya Pradesh, India

The present investigation aims to assess the processes controlling the composition of groundwater in the Sohagpur coalfield and its quality concerning drinking and irrigation purposes using hydrogeochemical and statistical approaches, water quality indices, and a geographic information system (GIS). Twenty-seven groundwater samples collected from different sites were analysed for electrical conductivity (EC), pH, turbidity, major anions and cations, total dissolved solids (TDS), and total hardness (TH). The study results show that the pH of samples varied from 5.97 to 8.26, suggesting the acidic to slightly alkaline nature of the water samples. The TDS ranged from 265 to 1450 mg/L with 78% of the samples being in the freshwater category. The Ca2+-Mg2+-SO4 2− and Ca2+-Mg2+-HCO3 − are dominant hydrogeochemical facies in the Sohagpur coalfield. The hydrogeochemical and statistical methods show that the groundwater chemistry of the Sohagpur coalfield is primarily controlled by the dissolution and weathering of minerals and secondarily due to the influence of anthropogenic activities (i.e., agricultural practices, sewage, and animal waste discharge). In the majority of the groundwater samples, concentrations of TH, TDS, turbidity, and SO4 2− exceeded the Bureau of Indian Standards (BIS) drinking acceptable limits and suggested that water is unsuitable for direct consumption. Moreover, the estimated water quality index (WQI) indicated that 52% of the samples (especially from the Dhanpuri and Amlai areas) belong to poor to unfit water categories. High values of EC in 56% of the samples and magnesium hazard (MH) at two sites restrict the irrigation suitability at these locations.

How wide is the problem? Leveraging alternative data sources to enhance channel width representation in watershed modeling

Watershed models have been increasingly applied to investigate the impacts of environmental changes on water quantity and quality and to support decision-making. Thus, accurately capturing physical processes and the characteristics of the watershed system is vital to achieving reliable results. Channel geometry features such as cross-sectional shape, bankfull channel width, and depth affect hydraulic parameters like hydraulic radius, wetted perimeter, and cross-sectional area. These, in turn, play important roles in streamflow dynamics, flow velocity, erosion, nutrient transport, and stream ecology. However, the representation of bankfull channel width is often oversimplified in watershed models. Power-law regression equations relating geometric parameters and upstream drainage area are commonly used to calculate bankfull channel width in watershed models. This approach has limitations and may resonate in the misrepresentation of channel geometry with implications for water quantity and quality estimations. This study evaluates how bankfull channel width is represented in a popular watershed model and presents alternative data sources derived from aerial measurements, empirical models, LiDAR, and a global database to mitigate potential misrepresentations. To assess the impacts of bankfull channel width on water quantity and quality, we designed a series of modeling experiments through the Soil and Water Assessment Tool (SWAT) model. We test our methodology in the Alabama-Coosa-Tallapoosa river basin, a large watershed system draining to the Gulf of Mexico coast. Our findings indicate that, overall, the regression equation used in the ArcSWAT program (SWAT's GIS interface) can overestimate bankfull channel width by as much as three times in our study domain. In testing the effects of bankfull width on model simulations we found negligeable implication for water yield predictions. On the other hand, 1-day maximum flows were greatly increased (20%) by using channel width values from the alternative data sources. Simulated water quality was also affected, with stream water temperature showing better agreement with observations under the proposed scenarios. Sediment loadings increased by as much as 118% under the alternative data sources of channel width. Nitrate and phosphate loadings had opposite responses and showed decreases as high as 8.3 and 18.8%, respectively, relative to the default model. Our findings demonstrate that bankfull channel width is misrepresented in watershed models with underlying impacts on simulated water quantity, quality, and ecological flows. Our analysis found bankfull channel width as an important parameter affecting high flows and water quality, and the freely available data sources tested in this study showed good matches with field measurements, making our findings readily available and broadly useful to the modeling community.

Water Quality Prediction Based on Hybrid Deep Learning Algorithm

Pollution from many different sources severely affects the quality of our water supply. Over the past few years, a large number of online water quality monitoring stations have been used to gather time series data on water quality monitoring. These numbers are the foundation for deep learning techniques for forecasting water quality. In particular, typical deep learning approaches struggle to accurately estimate water quality in the presence of net promoter system (NPS) contamination. To overcome this shortcoming, a new deep learning model called long short-term memory (LSTM)–gray wolf optimization (GWO)–fish swarm optimization (FSO) was developed to enhance the precision of water quality prediction with NPS pollution. The well-established model may remedy the mechanism models’ inability to foretell changes in water quality on a minute-by-minute basis. Thamirabarani river watershed was used for the model’s application. Based on experimental data, the suggested model outperformed the mechanism model and the LSTM model in predicting extreme values. Maximum relative errors in anticipated against observed dissolved oxygen, chemical oxygen demand, and NH3─N values were 7.58%, 18.45%, and 22.25%, respectively. In comparison to the artificial neural network (ANN), back propagation neural network (BPNN), and recurrent neural network (RNN) models, the created LSTM–GWO–FSO model was shown to have greater computational performance (RNN). LSTM–GWO–FSO outperformed ANN, BPNN, and RNN regarding R2 of 3.1%–38.4% improvements. The suggested approach may provide a fresh perspective when predicting water quality in the presence of NPS contamination.

Evaluation of water from Lake Coatetelco in central-south Mexico and surrounding groundwater wells for drinking and irrigation, and the possible health risks

Due to an increasing reduction of hydrological resources across Mexico and their growing contamination from global warming and anthropogenic activities, this study evaluated water from the perennial Lake Coatetelco (Ca–Mg–HCO3) in tropical central-southern Mexico and groundwater (Ca–Mg–HCO3 and Na–HCO3–Cl) from the surrounding wells for drinking as well as irrigation qualities. Comparison with the WHO guidelines and the estimated water quality indices (DWQI and IWQI) grouped almost all the samples collected after the warm season rainfall in excellent and good categories (DWQI < 100) for drinking, even though fluoride remained > 1.5 mg/L in 50% samples. Except for one groundwater sample, all showed > 25% permeability (classes I and II) in Donnen classification indicating their suitability for irrigation. USSL and Wilcox classifications, however, catalogued some in the high-salinity hazard group and some as doubtful for irrigating regular plants. Samples from about 53% wells were also in high and severe restriction categories of IWQI for the irrigation. Total Hazard Quotient Index (THQI) for estimating the non-carcinogenic risk (HQfluoride > 1) showed that at least one lake water sample and 53% of groundwater might expose the adult and child population to dental and skeletal fluorosis. This water quality assessment data posterior to the rainfall season could be useful as a baseline for both the short- and long-term monitoring in attention to the United Nation’s Sustainable Development Goal 6.

A2DWQPE: Adaptive and automated data-driven water quality parameter estimation

Accurate remote sensing estimation of inland water quality parameters (WQPs) plays a crucial role in guiding water resource management. To achieve this, researchers have explored various data-driven approaches utilizing machine learning (ML) techniques. However, there are two major challenges in WQPs estimation for inland waters. Firstly, current data-driven approaches focus on building a unified estimation model for an entire study area, which underestimates the complex dynamics of water constituents and optical properties. Secondly, ML models, particularly neural networks, require extensive hyperparameter tuning and are not user-friendly for researchers lacking relevant background and experience. In this paper, we propose an innovative method called adaptive and automated data-driven water quality parameter estimation (A2DWQPE) to address both challenges. Our method operates under the assumption that water bodies with similar spectral characteristics should share the same WQP estimation model. A2DWQPE is composed of three phases. Firstly, water types are automatedly classified by unsupervised hierarchical clustering according to spectral similarity. Then, optimal Deep Neural Network (DNN) models for estimating WQPs from multi-spectral satellite images are customized for each water type utilizing Bayesian optimization (BO). Finally, the target WQP is estimated based on the type-specific estimates and degree of membership of each water type. To evaluate the effectiveness of A2DWQPE, we applied it to estimate Secchi disk depth (SDD) in Lake Erie with in situ measurements and Moderate Resolution Imaging Spectroradiometer (MODIS) images. The results demonstrate that A2DWQPE outperforms the traditional approaches of developing a unified model for the entire study area. A2DWQPE achieved high accuracy with coefficient of determination (R2) over 0.72 and root mean square error (RMSE) below 1.4 m. Our method also outperforms the methods that applied Genetic Algorithm (GA) and Particle Swarm Optimization (PSO) instead of BO, and several traditional ML algorithms. We firmly believe that A2DWQPE holds great potential for accurate inland water quality estimation and will contribute significantly to various applications in water quality monitoring and pollution prevention.

Water Quality Estimation and Anomaly Detection: A Review

Critical infrastructures that provide irreplaceable services are systems that contain industrial control systems (ICS) that can cause great economic losses, security vulnerabilities and disruption of public order when the information in it is corrupted. These ICSs, which were previously isolated, have now become systems that contain online sensors, wireless networks and artificial intelligence technologies. This situation has also increased the scope of attacks by malicious people who intend to carry out industrial espionage and sabotage these systems. In this study, water quality estimation systems and anomaly detection are comprehensively examined. In this direction, the statistics of the studies in the literature, the methods for water quality anomaly detection, the existing data sets, and the difficulties encountered in the water systems to achieve better water management are discussed. Principle findings of this research can be summarized as follows: (i) new methodologies and architectures have improved water quality assessment through anomaly detection, (ii) different datasets including multi-modal information have been presented, and (iii) remaining challenges and prospects have been investigated.

Cyanobacterial bioreporter of nitrate bioavailability in aquatic ecosystems

The water eutrophication, resulting from the discharge of industrial and agricultural wastewater, leads to ecological degradation. However, to date, how to assess and manage the risks of water pollution, especially nitrogen pollution, remains a particularly noteworthy issue. Nitrate, the most important nitrogen compound, has become a bottleneck restricting total nitrogen management. The development of bioreporters monitoring nitrate pollution contributes to the estimation of water quality, especially the availability of nutrients. In this study, we obtained 9 bioreporters from 40 cyanobacterial derivatives which were constructed based on different hosts, copy numbers, and sensing elements and evaluated the performance of bioreporters. The results showed that single-celled Synechocystis was more sensitive to nitrate than filamentous Anabaena, that the reporter gene luxABCDE responded faster than sfgfp in most bioreporters, and that relatively medium-copy plasmid improved the performance of sensing elements. Nine bioreporters performed well in bioavailable nitrate detection, of which AD-AS-X and AR-NI-X, activated by nitrate repletion, had the shortest response time (2 h) and the widest response range (20–800 μM), respectively. Moreover, SR-GLN-SG, activated by nitrate deficiency, exhibited the best linear response (R2 = 0.998). After parameter optimization, exponential growth phase bioreporters, culture temperature of 30 °C, sample volume of 200 μL were determined as optimal monitoring conditions. We found that common water contaminants (copper, cadmium, and phosphorus) had no impact on the performance of bioreporters, indicating the stability of bioreporters. Six out of 9 bioreporters, especially the SR-NB-X, were highly effective in detecting the bioavailable nitrate in wastewater sample. This study provides valuable references for developing more cyanobacterial bioreporters and their practical application in nitrate detection.

A Machine Learning-Based Framework for Water Quality Index Estimation in the Southern Bug River

River water quality is of utmost importance because the river is not only one of the key water resources but also a natural habitat serving its surrounding environment. In a bid to address whether it has a qualified quality, various analytics are required to be considered, but it is challenging to measure all of them frequently along a river reach. Therefore, estimating water quality index (WQI) incorporating several weighted analytics is a useful approach to assess water quality in rivers. This study explored applications of ten machine learning (ML) models to estimate WQI for the Southern Bug River, which is the second-longest river in Ukraine. The ML methods considered in this study include artificial neural networks (ANNs), Support Vector Regressor (SVR), Extreme Learning Machine, Decision Tree Regressor, random forest, AdaBoost (AB), Gradient Boosting Regressor, XGBoost Regressor (XGBR), Gaussian process (GP), and K-nearest neighbors (KNN). Each data measurement consists of nine analytics (NH4, BOD5, suspended solids, DO, NO3, NO2, SO4, PO4, Cl), while the quantity of data is more than 2700 data points. The results indicated that all ML models demonstrate satisfactory performance in predicting WQI. However, GP outperformed the other models, followed by XGBR, SVR, and KNN. Furthermore, ANN and AB demonstrated relatively weaker performance. Moreover, a reliability assessment conducted on both training and testing datasets also confirmed the results of the comparative analysis. Overall, the results enhance the assertion that ML models can sufficiently predict WQI, thereby enhancing water quality management.

Estimation of water quality of Dikhu River of Nagaland through a combination of water quality index and principal component analysis techniques.

Estimation of water quality of Dikhu River of Nagaland through a combination of water quality index and principal component analysis techniques.

Measurement of Turbidity and Total Suspended Matter in the Albufera of Valencia Lagoon (Spain) Using Sentinel-2 Images

In the context of freshwater ecosystems, turbidity and suspended solids play crucial roles, with their levels significantly influenced by anthropogenic activities. This study focuses on assessing and monitoring these parameters in Albufera de Valencia using Sentinel-2 imagery. The primary aim is to establish reliable estimation algorithms that can contribute to effective ecosystem management. This study calibrated and validated algorithms for estimating turbidity and suspended solids. The R783 × R705/R490 model proved to have the best performance for estimating turbidity and total solids in Albufera. However, R783/R490 obtained a higher coefficient of determination for the organic part, while the R705 band was selected for the inorganic part. However, to achieve better estimates of turbidity and inorganic matter, more research is needed in the future. The implications of excessive suspended solids are underscored, including the depletion of dissolved oxygen and reduced primary productivity due to limited light penetration and habitat availability. Collaboration between disciplines such as limnology, optics, and water chemistry is crucial to advance water quality estimation models in lakes and lagoons such as Albufera. By integrating expertise and approaches from these diverse fields, new knowledge can be gained, and the basis for more effective management and conservation strategies can be laid.

Environmental Consequences of Intensive Aquaculture Practices at Moyna Purba Medinipur West Bengal India with Special Reference to Carbon Footprint and Carbon Sequestration.

Moyna of Purba Medinipur district is widely known as an aquaculture hub of West Bengal, India. Though very good production is achievable from this fish culture system, management practices are inappropriate, which could create the sustainability problem of this culture system. The present study was thus undertaken for the estimation of plankton population, water quality, carbon footprint and carbon sequestration of this intensive aquaculture practices. Information on spawn to fry, fry to fingerlings and grow-out culture were collected through the structured questionnaire from the fish farmers. The plankton density, primary productivity, carbon footprint and carbon sequestration were analyzed using standard procedures. The phytoplankton, zooplankton and primary productivity were maximum at the stocking period and minimum during the middle of culture period. The lowest feed conversion ratio (FCR) was noticed with the minimum amount of feed applied in the pond. The CO2-e emission ranged from 0.56 to 4.89kg CO2-e/kg fish (av. 2.13) for the production levels of 5.0 to 10.7 t/ha/yr. The pond water developed salinity and ammonium-N increased from 0.01 to 0.50mg/l. The ponds with high feed loading (28 to 32 t/ha/yr) had the highest average sediment accumulation rate (11.0 ± 3.0cm/yr) and carbon sequestration (704 ± 30g C/m2/yr).

Monitoring multi-water quality of internationally important karst wetland through deep learning, multi-sensor and multi-platform remote sensing images: A case study of Guilin, China

Karst wetlands are widely distributed throughout the southwest China, and play an important role in enhancing carbon sequestration and improving water quality in karst areas. The internationally important karst wetland of Huixian is the largest karst wetland in China, but its water quality has continued to deteriorate as a result of human influences in recent years. Remote sensing technology has become an important approach to estimate water quality parameters (WQPs). However, the feasibility of combining multi-sensor remote sensing images with deep learning to estimate different WQPs in karst wetlands has not been demonstrated yet. To resolve this issue, this study constructed multiple retrieval models of WQPs (Chlorophyll-a (Chla), Phycocyanin (PC), Turbidity (Turb), Dissolved Oxygen (DO)) in karst wetlands using deep learning (Transformer and Mixture Density Network (MDN)) and optimized shallow machine learning (Random Forest (RF), XGBoost (XGB) and Gradient Boosting (GB)) based on multi-sensor images from satellite and UAV platforms. The performance of deep learning in the inversion of WQPs demonstrated to compare with shallow machine learning using multispectral and hyperspectral images. We further quantitatively evaluated the retrieval performance of UAV and satellite, multispectral and hyperspectral images, and presented predictive mapping of the gradient distribution of WQPs. Finally, this study adopted the SHapley Additive exPlanations (SHAP) to tackle the local and global interpretability of the input features contribution to the output of retrieval models. The results showed that (1) Transformer model presented a good prediction of PC and DO (R2 = 0.649 ∼ 0.844), XGB and GB models achieved the highest accuracy estimation of Chla and Turb (R2 = 0.75). (2) The estimation results of WQPs based on UAV platform (R2 = 0.419 ∼ 0.695) was higher than that of satellite-based images. The estimation accuracy of multispectral images (R2 = 0.338 ∼ 0.718) was slightly higher than that of Zhuhai-1 Orbita hyperspectral (OHS) images. The average accuracy of Turb estimated by UAV images (R2 = 0.565 ∼ 0.752) was higher than that of satellite-based images. OHS hyperspectral images had the best DO estimation (R2 = 0.314 ∼ 0.649). (3) This study found 32.66% and 23.01% of water area with the Chla and Turb concentrations exceeding 60 μg/L and 60 NTU, respectively, which revealed that the Huixian karst wetland has suffered serious water pollution. (4) The SHAP analysis reveals that near infra-red and red band are sensitive to predict Chla and DO, red and red-edge bands are sensitive to predict PC and Turb in the karst wetland.