ABSTRACT
This study's objective was to evaluate groundwater geochemical processes, investigate the extent and sources of heavy metals in soils, and develop models for predicting PM2.5 within Iyuku and Ikpeshi and environs. A total of 32 groundwater samples, in addition, data on static water level and drill depths of two boreholes in Ikpeshi were obtained from the Benin Owena River Development Authority (BORDA), while the static water levels and mean sea elevations of 8 hand dug wells and mean sea elevations were determined. 120 soil samples from 30 sampling points including control at Agbede were collected in two (2) depths; topsoil (0-15 cm) and subsurface soil (15-30 cm). Information on temperature, relative humidity level (RH), wind speed (WS) and PM2.5 in the air were randomly sampled in nine locations. Kriging and interpolation techniques were used to produce geospatial distribution map of PM2.5. The heavy metal parameters (Fe, Zn, Cu, Pb, Mn, Ni, Cd, Cr and Co) in groundwater and soil samples were analyzed with Atomic Absorption Spectrometer (AAS VGP210 Bulk Scientific). Ion Chromatographic method and titrimetric method was used to determine HCO3, Cl, SO4, NO3, and OM.
Piper, Gibb’s, and ionic plots delineated hydrogeochemical facies as CaCl water type and groundwater quality in the area is predominantly influenced by disintegration (weathering) of rocks (quarzite schist and granites) in the study area. Fe demonstrated a strong positive association with Zn, Cu, Mn, and Ni, and a positive correlation with TDS, Cl, and HCO3. Groundwater flow model indicated that the direction of groundwater flow is towards the southwestern region of the study area. Soil samples demonstrated a slightly acidic pH of 6.79 and 6.84 in the topsoil and subsurface soil respectively. It was observed that the metals (Fe, Zn, Cd, Mn, Ni, Co and Cr) in soil sampling locations were significantly higher compared to Control (p<0.05). Result of Enrichment factor and Igeo showed moderate enrichment for (Mn) and (Ni) at topsoil and linked their origin to anthropogenic sources. The result of ecological risk index shows moderate to considerable contamination by cadmium (Cd) and manganese (Mn). (MCD) and (PLI) ranged from 3.29 to 4.06 and 4.37 to 4.28 in topsoil and subsoil indicating moderate to severe pollution. It was noted that iron (Fe), zinc (Zn) and copper (Cu) at topsoil and subsurface soil exhibited a strong positive correlation with Electrical conductivity (EC). Fe, Cu and Zn exhibited substantial negative correlation with pH. Results of cluster analysis highlighted the existence of two dominant relationships between Cu, Zn and Fe and Zn and EC in top soil and Cu, Zn and Fe and OM and EC in subsoils. Principal component analysis reveals three (3) significant PCs for topsoil (>0.7) PC1(Fe, Zn, Cu, and Mn), PC2(Co) and PC3 (OC and OM). At subsoil only 2 PCs were significant (>0.7), PC1(Fe, Zn, Pb, Cd, Mn and EC), PC2 (OM and OC).
Results of analysis conducted on air quality parameters (PM2.5, windspeed, relative humidity and Temperature) were as follows. Results of Normality (Jarque - Bera value ≤10 and P<0.05) suggests all variables except humidity did not obey normality. Homogeneity Test reveals that all variables were statistically homogenous. Outlier Test using the 25th and 75th percentile indicated the absence of outliers in the data. Crochbach alpha value of 0.83 (83 %) and p<0.05 attested to the reliability and significance of the data. Findings from result of forecasting models using MATLAB R2019a reveals that the predictive abilities of Multiple linear regression models (MLR) was good with R2= 0.8911. However, with standard error (SE) of 12.40, it may likely be inadequate in predicting PM2.5. It was observed that artificial neural network (ANN) has a better predictive performance with R2= 0.9373 and low mean square error (MSE)= 0.000158 at epoch 10. Spatial distribution map of PM2.5 within the study area produced using kriging and interpolation designated the Northwestern part of the study area as hotspots of PM2.5 pollutions. It can therefore be concluded that this research has opened a new approach towards understanding geochemical evolution of groundwater, source apportionment/ identification of contaminantsinsoil.
