Application of Correlation Matrix and Analysis of Variance (Anova) in Profiling Surface Water in Nembe Creek of Niger Delta, Nigeria

Abstract
Water resource contamination is a major concern in several regions, especially in the Niger Delta, in which the oil exploration activities pose serious risks to human health, water resource and the environment. This present study evaluates the properties of surface water in Nembe Creek. Surface water samples were collected from three (3) control stations and twenty-four (24) sampling stations around the creek during the wet and dry seasons. The samples were subjected to analysis following the standard operating procedures of ASTM and APHA analytical methods. The relationship between the parameters and variations of surface water quality of Nembe creek were determined by using descriptive statistics, Analysis of variance (ANOVA) and Pearson correlation analysis. The results indicated pH had a mean value of 6.67± 0.33 (wet season) and 7.21 ± 0.3 (dry season), Electrical Conductivity ranged 200 – 3950 µS/cm (wet season) and 17900 – 25800 µS/cm (dry season), DO ranged between 4.3 – 5.8 (wet season) and 3.9 – 5.6 (dry season), Nitrate had values of 0.007- 0.068 mg/L (wet season) and 0 – 0.007 mg/L (dry season), Sulphate values range 130.58- 158.76 mg/L (wet season) and 215.12 – 657.35 mg/L (dry season). Iron range 3.23 – 5.6 mg/L (wet season) and 0.26 – 7.87 mg/L (dry season), Zinc ranged 0.02 – 0.47 mg/L (wet season) and 0.13 – 2.22 mg/L (dry season). Across all sampling stations and control areas during the wet season, the values for Total Petroleum Hydrocarbon (TPH), total hydrocarbon Content (THC) and Petroleum Aromatic Hydrocarbon (PAH) were below the detection limit; however, the values varied during the dry season. The Pearson Correlation matrix revealed positive, negative and significant correlations between the pairs of parameters at 95% confidence level. Amongst the sampling and control stations, some parameters had no significant difference (p > 0.05), however some differences were statistically significant (p < 0.05). In conclusion, Nembe Creek is marginally more polluted during the dry season than during the wet season. Most parameters on the Person's Correlation matrix were connected with each other based on the metrics. However, some parameters such as metals, do not significantly correlate with one another, indicating that the sources of contamination may have originated from various places. It is possible that the source may not predominantly originate from oil exploration activities but may involve other natural, geological and anthropogenic activities.