MODELING OF ANISOTROPIC SEMIVARIOGRAM USING WEIGHTED LEAST SQUARE AS PARAMETER ESTIMATION METHOD (CASE STUDY: HEAVY METAL DATA IN WELL WATER ON BANDUNG REGENCY)

• Main Author: Lia Putri, Rina
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/84071
• Institution: Institut Teknologi Bandung

Well water is an essential resource for the residents of Bandung Regency. However, the presence of heavy metals such as lead (Pb) and cadmium (Cd) in concentrations exceeding regulatory limits poses a significant health risk. According to Government Regulation No. 82 of 2001, the maximum permissible concentrations of Cd and Pb in water are 0.01 mg/L and 0.03 mg/L, respectively. Spatial modeling of heavy metal concentrations in well water can be effectively conducted using the geostatistical method of semivariogram analysis. A semivariogram quantifies the spatial correlation between sampled locations. In this study, parameters such as the nugget effect (????0), sill (????0+????), and range (????) were estimated using the Weighted Least Square (WLS) method, an advancement of the Least Squares (LS) method. The optimal semivariogram model was then utilized to interpolate unobserved locations through Ordinary Kriging interpolation. The data comprised Cd and Pb concentrations in well water across Bandung Regency, modeled using Gaussian, Exponential, Spherical, and Cubic theoretical semivariogram models. The findings indicate that the anisotropic semivariogram model representing Cd is best characterized by a Zonal Cubic model, exhibiting a minimum sill in the East-West direction and a maximum sill in the North-South direction, with a WLS SSE of 786.98. Similarly, the Pb distribution is best represented by a Zonal Cubic model, with a minimum sill in the Southeast-Northwest direction and a maximum sill in the North-South direction, with a WLS SSE of 1,120.98. The results of the Ordinary Kriging interpolation illustrate the spatial distribution of Cd and Pb concentrations at unsampled locations.