Statistical Modeling of the Groundwater Arsenic Contamination Level in Bangladesh Due to Chemical Elements

Syed Md Fakrul AHSAN
Md. Nazrul ISLAM
Md. Jamal UDDIN
Mohammed Taj UDDIN

Keywords

regression model, principal component, spatial autocorrelation, arsenic contamination, chemical elements

Abstract

This paper is devoted to modeling of the magnitude of arsenic with 54 variables of chemical elements based on secondary data. The samples are collected from 113 wells from different areas of Bangladesh. The study shows that 9 variables are significantly related with arsenic variable. Analysis also claims that arsenic variable can be modeled with three principal components (linear combination of independent variables). The study also considered latitude and longitude variables but there is no spatial autocorrelation between them.