TY - JOUR
T1 - The use of statistical methods to assess groundwater contamination in the Lower Tano river basin, Ghana, West Africa
AU - Edjah, Adwoba K.M.
AU - Banoeng-Yakubo, Bruce
AU - Akiti, Thomas Tetteh
AU - Doku-Amponsah, Kwabena
AU - Duah, Anthony A.
AU - Sakyi-Yeboah, Enoch
AU - Kippo, Joseph Valcory
AU - Amadu, Inusah
AU - Ibrahim, Kwabina
N1 - Publisher Copyright:
© 2021, The Author(s), under exclusive licence to Springer Nature Switzerland AG.
PY - 2021/11
Y1 - 2021/11
N2 - In this study, descriptive statistics, correlation matrix, multiple regression model, and geostatistical models were used to assess the contamination of groundwater with respect to trace elements in the Lower Tano river basin, Ghana, West Africa. A total number of 48 boreholes drilled across the basin with depths ranging from 18 to 60 m were used as data sources in this study. The results of the descriptive statistics showed that the average lead, iron, and aluminium concentrations exceeded the WHO permissible limits of 0.3 mg/L, 0.01 mg/L, and 0.2 mg/L respectively. Furthermore, copper, chromium, aluminium, zinc, manganese, nickel, iron, arsenic, electrical conductivity, and total dissolved solids were found to be extreme and highly positively skewed. Even though significant correlations exist among some variables, the statistical results showed that the quality of the boreholes drilled across the basin was mainly originating from geogenic and anthropogenic sources. In addition, each pair of correlated physical parameters and trace elements in the drilled boreholes were predicted using multiple regression models. Likewise, geostatistical modelling was used to assess the spatial analysis of each pair of correlated physical parameters and trace elements in the drilled boreholes. The cross-validation results revealed kriging model, as the most precise model for the spatial distribution maps for the correlated physical parameters, and correlated trace elements concentration in the boreholes drilled across the study region. The semivariogram models showed that most of the correlated physical parameters and correlated trace elements were weak moderately and strongly spatially dependent, suggesting fewer agronomic influences. The results of the spatial analysis were consistent with the multiple regression model and the Pearson correlation matrix.
AB - In this study, descriptive statistics, correlation matrix, multiple regression model, and geostatistical models were used to assess the contamination of groundwater with respect to trace elements in the Lower Tano river basin, Ghana, West Africa. A total number of 48 boreholes drilled across the basin with depths ranging from 18 to 60 m were used as data sources in this study. The results of the descriptive statistics showed that the average lead, iron, and aluminium concentrations exceeded the WHO permissible limits of 0.3 mg/L, 0.01 mg/L, and 0.2 mg/L respectively. Furthermore, copper, chromium, aluminium, zinc, manganese, nickel, iron, arsenic, electrical conductivity, and total dissolved solids were found to be extreme and highly positively skewed. Even though significant correlations exist among some variables, the statistical results showed that the quality of the boreholes drilled across the basin was mainly originating from geogenic and anthropogenic sources. In addition, each pair of correlated physical parameters and trace elements in the drilled boreholes were predicted using multiple regression models. Likewise, geostatistical modelling was used to assess the spatial analysis of each pair of correlated physical parameters and trace elements in the drilled boreholes. The cross-validation results revealed kriging model, as the most precise model for the spatial distribution maps for the correlated physical parameters, and correlated trace elements concentration in the boreholes drilled across the study region. The semivariogram models showed that most of the correlated physical parameters and correlated trace elements were weak moderately and strongly spatially dependent, suggesting fewer agronomic influences. The results of the spatial analysis were consistent with the multiple regression model and the Pearson correlation matrix.
KW - Descriptive statistics
KW - Geostatistical models
KW - Groundwater contamination
KW - Lower Tano river basin
KW - Multiple regression model
UR - http://www.scopus.com/inward/record.url?scp=85117865787&partnerID=8YFLogxK
U2 - 10.1007/s10661-021-09514-z
DO - 10.1007/s10661-021-09514-z
M3 - Article
C2 - 34694510
AN - SCOPUS:85117865787
SN - 0167-6369
VL - 193
JO - Environmental Monitoring and Assessment
JF - Environmental Monitoring and Assessment
IS - 11
M1 - 748
ER -