Multiple linear regression analysis (MLR) applied for modeling a new WQI equation for monitoring the water quality of Mirim Lagoon, in the state of Rio Grande do Sul—Brazil

Abstract

Accurate assessment of the type and extent of water pollution is a difficult and complicated task. Therefore, the use of the Water Quality Index (WQI) proves to be a useful tool, as this index has the advantage of resulting in a single number that is easy to communicate and understand. One of the statistical methods that can be used to develop a new WQI equation for a given water body is Multiple Linear Regression (MLR). Therefore, this work aims to develop a new WQI equation for Mirim Lagoon through MLR and to evaluate the suitability of that equation. For this, 154 samples collected at 7 monitoring points of Mirim Lagoon over three years were analyzed, from which the original WQI results were obtained. With these results, MLR was applied, generating a new WQI equation with only three variables: phosphorus, dissolved oxygen (DO) and thermotolerant coliforms. This MLR model obtained an R2 coefficient equal to 0.728, demonstrating that 72.8% of the data variability was explained by the equation generated by this model. Through the paired t test it was possible to demonstrate that the results of the new WQI did not differ significantly from the results of the original WQI, obtaining significance values greater than 0.05 for both absolute WQI values and quality ranges (p value 0.886 and 0.467, respectively), attesting to the adequacy of the new equation. The reduction in the number of variables for monitoring water quality has several advantages, from financial ones, as it makes monitoring less costly, even in reducing the eclipse effect. Therefore, it is concluded that it is of great importance to generate an equation with fewer variables for monitoring the quality of the waters of Mirim Lagoon.

Performance of machine learning methods in predicting water quality index based on irregular data set: application on Illizi region (Algerian southeast)

“…Recently, Singha et al (2021) applied deep learning for predicting WQI with 3 traditional models and found that the deep learning model is a more robust and accurate tool than the traditional model in the prediction of groundwater quality. Valentini et al (2021) introduced a new WQI equation for Mirim Lagoon and evaluated its suitability based on 154 samples collected over three years at seven sampling points in Mirim Lagoon. For forecasting monthly WQI values at the Lam Tsuen River in Hong Kong, Asadollah et al (2021) proposed a new ensemble machine learning algorithm called extra tree regression (ETR).…”

Hybrid Technique to Improve the River Water Level Forecasting Using Artificial Neural Network‐Based Marine Predators Algorithm

“…The value SI � 0.0009 means that the MPA-ANN has an excellent performance according to the limitations in Section 3.4. Furthermore, the Kolmogorov-Smirnov and Shapiro-Wilk tests agree that the residual data are normally distributed depending on the significant values (p value) being more than 0.05 [72,73]. The study's statistical testing findings show the following: (1) Data preprocessing techniques play an essential role in improving data quality, particularly the SSA method.…”

Analysis of the Influence of Natural Factors on the Bathing Condition of Beaches in the Cities of Garopaba, Imbituba and Laguna, in the State of Santa Catarina – Brazil

“…The use of nonparametric tests, such as those used in this study, to assess the assumptions of sample distribution of data is something quite consolidated in the literature and several authors have used these tests in studies related to water quality. Furthermore, the KS and SW normality tests are applied in several studies regarding surface or groundwater quality (e.g., Santos et al, 2020;Batista et al, 2021;Leite et al, 2021;Valentini et al, 2021c;Valentini et al, 2021d;Silveira et al, 2021). Santos et al (2021), for example, applied KS and SW tests for evaluate data sampling distribution of in order to know which correlation coefficient would be more appropriate for application of correlation analysis between these data and Water Quality Index calculated in their study.…”