Monte Carlo Uncertainty Analysis and Regression Models on Water Quality Estimation Strait of Tuba, Langkawi

  • Hasmida Muhamad Faculty of Architecture, Planning and Surveying, Universiti Teknologi MARA, Perlis Branch, Arau Campus, 02600 Arau, Perlis, Malaysia
  • Ernieza Suhana Mokhtar Faculty of Architecture, Planning and Surveying, Universiti Teknologi MARA, Perlis Branch, Arau Campus, 02600 Arau, Perlis, Malaysia
  • Muhammad Akmal Roslani Faculty of Applied Sciences, Universiti Teknologi MARA, Perlis Branch, Arau Campus, 02600 Arau, Perlis, Malaysia
  • Idrees Mohammed Oludare
  • Azlan Abdul Aziz
  • Noraini Nasirun

Abstract

Preserving and maintaining water quality is essential in providing sustainable development for a nation, particularly for the protection of marine life and human health. In acquiring primary data for water quality parameters (WQP), it is indisputable how prominent the in-situ sampling technique is as opposed to other methods (e.g., satellite imagery, drones). However, the unpredictable condition of the environment and other uncertainties that may exist in the acquired data from in-situ sampling needs to be assessed to ensure its consistency and credibility. Hence, this research aims to identify the best linear regression model for estimating WQP in the Strait of Tuba, Langkawi by i) reducing the existing error in sampling data using Monte Carlo (MC) analysis and ii) determining the best spatial interpolation method to interpolate the WQP. About 71 sampling points were collected in December 2021 situated in Selat River, Langkawi. An uncertainty MC analysis was applied to the sampling data and the mean distribution from the MC analysis result has less error compared to the observed data. The result for the standard deviation of turbidity, salinity, temperature, pH, and dissolved oxygen (DO) are 0.2333, 0.6695, 0.4711, 0.2671, and 1.3230 respectively. For the spatial interpolation model comparison, Inverse Distance Weightage (IDW) spatial interpolation model outperforms kriging in terms of RMSE, Mean Absolute Error (MAE), and standard deviation. Finally, nine linear regression equations were established to predict the WQP. This study will help decision-makers in predicting the WQ by reducing the ground sampling data collection with the proposed linear regressions.


Keywords: Water Quality, Geospatial Analysis, Spatial Interpolation, uncertainty analysis, regression model

Published
2022-07-03
How to Cite
MUHAMAD, Hasmida et al. Monte Carlo Uncertainty Analysis and Regression Models on Water Quality Estimation Strait of Tuba, Langkawi. Built Environment Journal, [S.l.], v. 19, n. 2, p. 129-137, july 2022. ISSN 2637-0395. Available at: <https://myjms.mohe.gov.my/index.php/bej/article/view/17849>. Date accessed: 21 july 2024. doi: https://doi.org/10.24191/bej.v19i2.17849.