Evaluation of Water Quality Based on a Machine Learning Algorithm and Water Quality Index for Mid Gangetic Region (South Bihar plain), India

Abstract Water is an indispensable resource for human production and life. The evaluation of water quality by scientific method that provides sufficient support for the regeneration and recycling utilization of water resources. At present, water quality is mainly evaluated by water quality index (WQI) with weighted entropy value, which comprehensively considers the influence of different relevant environmental factors on the water quality. The calculation process is very complicated and time-consuming. In this paper, the method of correlation analysis is used to select the best combination of relevant environmental factors to assist the prediction model. Two typical kinds of machine learning methods are adopted and compared to realize the prediction of entropy water quality index (EWQI). After the better framework of prediction model is selected, four different kinds of optimization algorithms are used to optimize the prediction model to realize non-linear regression prediction and classification of water quality. According to the results of evaluation indicators, the framework of SVM is more suitable for realizing the prediction of EWQI. Meanwhile, the optimization algorithm of DE-GWO show great potential to improve the performance of SVM, which can make further contribution to the rational use and protection of water resources.

Download Full-text

Prediction of groundwater quality indices using machine learning algorithms

Water Practice & Technology ◽

10.2166/wpt.2021.120 ◽

2021 ◽

Author(s):

Hemant Raheja ◽

Arun Goel ◽

Mahesh Pal

Keyword(s):

Machine Learning ◽

Water Quality ◽

Water Quality Index ◽

Quality Index ◽

Learning Algorithms ◽

Machine Learning Algorithms ◽

Gradient Boosting ◽

Quality Indices ◽

Extreme Gradient Boosting ◽

Analysis Of Results

Abstract The present paper deals with performance evaluation of application of three machine learning algorithms such as Deep neural network (DNN), Gradient boosting machine (GBM) and Extreme gradient boosting (XGBoost) to evaluate the ground water indices over a study area of Haryana state (India). To investigate the applicability of these models, two water quality indices namely Entropy Water Quality Index (EWQI) and Water Quality Index (WQI) are employed in the present study. Analysis of results demonstrated that DNN has exhibited comparatively lower error values and it performed better in the prediction of both indices i.e. EWQI and WQI. The values of Correlation Coefficient (CC = 0.989), Root Mean Square Error (RMSE = 0.037), Nash–Sutcliffe efficiency (NSE = 0.995), Index of agreement (d = 0.999) for EWQI and CC = 0.975, RMSE = 0.055, NSE = 0.991, d = 0.998 for WQI have been obtained. From variable importance of input parameters, the Electrical conductivity (EC) was observed to be most significant and ‘pH’ was least significant parameter in predictions of EWQI and WQI using these three models. It is envisaged that the results of study can be used to righteously predict EWQI and WQI of groundwater to decide its potability.

Download Full-text

Developing a Machine Learning Algorithm to Identify Hot Moments Using Mass Balance Approach at Watershed Scale

10.5194/egusphere-egu21-16395 ◽

2021 ◽

Author(s):

Thakshajini Thaasan ◽

Phung Quang ◽

Noel Aloysius

Keyword(s):

Machine Learning ◽

Water Quality ◽

Mass Balance ◽

Crop Yield ◽

Land Management ◽

Learning Algorithm ◽

Machine Learning Algorithm ◽

Watershed Scale ◽

Relative Change ◽

Mass Balance Approach

Preserving and promoting the sustainable use of natural resources while stabilizing healthy ecosystems under rapid environmental changes is a tremendous challenge for the international community. Science-based strategies are imperative to maintain and improve Earth&#8217;s ecosystem. Our research is designed to improve predictive ability of managed ecosystems&#8217; responses to changing weather patterns and human management. Specifically, our research seeks to develop conservation plans to improve water quality in streams and lakes, while maintaining the economic sustainability of food production systems. Reducing pollution loading into aquatic systems help improve the water quality and enhance ecosystem sustainability. Non-point pollution sources are predominant factors in increasing pollution into the water bodies. Identifying the pollution sources is important to mitigate the impact. For this reason, the main objective of our study is to identify the &#8220;hot spots&#8221; and &#8220;hot moments&#8221; of excessive nitrogen and phosphorus leaching from managed landscapes in the midwestern United States.We developed a simple lumped model with three parameters to simulate key water fluxes - surface and subsurface runoff, and evapotranspiration (ET) in the Maumee River Basin. We designed a machine learning algorithm to identify &#8220;hot moments&#8221; using nitrogen mass balance approach at watershed-scale. The simple model helps to link the relationship between applied fertilizer and retained nutrients in the soil that the heterogeneous landscape and land management influence. Nitrogen retained in the soil will be used as an output variable and connected with predictor variable ET. Relationships between crop yield and water use in crop growth (ET) could be interpreted in a simple empirical formulation where relative change in crop yield is related to the corresponding relative change in ET, which can be expressed as,1&#8722;&#119884;&#119886;/ &#119884;&#119909;=&#119870;&#119910; (1&#8722; &#119864;&#119879;&#119909;/&#119864;&#119879;&#119886;)where Yx and Ya are the maximum and actual yields, ETx and ETa are the maximum and actual evapotranspiration, and Ky is a yield response factor representing the effect of relative change in ET on crop yield. The developed algorithm will be trained, tested, and validated using the coupled water flux and crop yield models. We will then demonstrate how these relationships can be extended to complex watershed model simulations that account for key land management decisions, land use pattern, crop type, soil, and topographic variability. Ultimately, we hope our findings will enhance the knowledge related to the environmental policy and decision making.

Download Full-text