References
1 Oil-spilled communities, Delta State 0.023 200.1 Audu et al., (2019) 2 School of Marine Technology, Burutu Delta State 0.015 130.5 Esi et al., (2019) 3 Central Part of Delta State 0.010 87.0 Edomi et al., (2020) 4 Fuel/Gas Dispensing Stations, Warri Metropolis 0.015 130.5 Agbalagba et al., (2020) 5 Lead/Zinc Mining Sites, Ishiagu, Ebonyi State 0.017 147.9 Mgbeokwere et al., (2021) 6 Nembe Clusters Oil/Gas Areas, Bayelsa State 0.026 226.2 Esendu et al., (2021) 7 Solid mineral mining locations, Edo-North Nigeria 0.021 182.7 Avwiri et al., (2021) 8 Kolo, Bayelsa State 0.018 156.6 Anekwe et al., (2023) 9 Beta Glass Plc, Ughelli 0.010 87.0 Onyekachi et al., (2023) 10 Solid mineral mining sites, Ebonyi State 0.020 174.0 Mgbeokwere et al., (2023) 11 Okutukutu Computer Village, Yenagoa, Bayelsa State 0.016 139.2 Biere et al., (2025) 12 FUPRE and PTI Campus, Effurun, Delta State 0.012 104.4 Present Study 3.3 Spatial and Statistical Analysis of BIR The statistical distribution of background ionizing radiation values was first examined using the Anderson–Darling test and Quantile–Quantile (Q–Q) plots. Both campuses exhibited slight departures from normality, with FUPRE recording an Anderson–Darling statistic of 0.5877 and PTI 0.7769. The Q–Q plots (Figure 3) confirm mild deviations from the expected straight line, particularly at the distribution tails, reflecting heterogeneous environmental sources and localized influences such as construction materials and micro-geological variation. Figure 3: Q–Q plot of BIR distribution in FUPRE and PTI Campuses Spatial autocorrelation analysis using Global Moran’s I revealed weak positive clustering of BIR values, though not statistically significant under permutation testing (FUPRE: I = 0.1737, z = 1.729, p = 0.0610; PTI: I = 0.0839, z = 1.849, p = 0.0610). These results suggest near-random distribution of radiation levels at the sampled resolution, with only modest local reinforcement. Detailed Moran scatterplots are provided in the Supplementary Document for robustness. Although Moran’s I values indicate weak spatial autocorrelation, the presence of localized hotspots in the kriging maps (figure 4 and 5) suggests that radiation distribution is influenced by site- specific environmental and anthropogenic factors rather than broad spatial dependence. The variogram modeling and cross-validation identified the spherical model as optimal for FUPRE (RMSE = 0.005056; MAE = 0.004517) and the linear model for PTI (RMSE = 0.003258; MAE = 0.002765). These models were subsequently used for ordinary kriging interpolation to generate continuous exposure. Figure 4: Krigged prediction map of BIR in FUPRE Campus (variogram=spherical) Figure 5: Krigged prediction map of BIR in PTI Campus (variogram = linear) Model validation through Leave-One-Out Cross Validation confirmed adequate predictive skill, with observed–predicted scatter clustered around the 1:1 line. Error magnitudes were proportionate to campus variability, with lower errors at PTI reflecting its narrower exposure range. Supplementary figures provide full LOOCV plots. The uncertainty in dose estimates was further quantified using Monte Carlo simulation (10,000 iterations, lognormal distribution, 5% relative uncertainty). The resulting prediction intervals (figure 6 and 7) confirmed the robustness of kriging outputs, with FUPRE exhibiting slightly higher uncertainty bands than PTI, consistent with its broader exposure range and localized hotspots. Figure 6: Monte-Carlo simulation output (FUPRE Campus) Figure 7: Monte-Carlo simulation output (PTI Campus) Unlike conventional reporting of background ionizing radiation that emphasizes absorbed dose rates or annual effective dose equivalents, this study applied the Radiological Risk Index as an integrative classification tool. The RRI combines AEDE and ELCR values and categorizes exposure points against UNSCEAR and ICRP thresholds. The strength of RRI lies in translating quantitative dose metrics into qualitative categories (very low, low, moderate), thereby enhancing interpretability for environmental health assessments in settings influenced by construction activities, geological heterogeneity, or anthropogenic inputs. Application of the RRI showed that FUPRE contained a mix of categories: very low (3 points), low (11 points), and moderate (6 points), with moderate risk zones concentrated in areas of elevated BIR, consistent with hotspots identified in kriging maps and plausibly linked to recent construction. In contrast, PTI was dominated by very low risk (13 points) and low risk (7 points), reflecting its lower exposure levels and more uniform distribution. The RRI maps and pointwise tables are provided in the Supplementary Document. 4.0 Conclusion and Recommendations The present study has demonstrated that background ionizing radiation levels across the two campuses investigated remain largely within internationally accepted global average values, though with notable differences between sites. FUPRE exhibited marginally elevated exposure rates, with localized hotspots that slightly exceeded UNSCEAR reference values. These elevations are plausibly linked to recent construction activities and the use of building materials containing trace radionuclides. PTI, by contrast, maintained consistently lower exposure levels, well below recommended thresholds, reflecting a more uniform and radiologically safe environment. Dose assessments confirmed that exposures across both campuses remain below the annual public dose limit of 1 mSv y−1, while spatial–statistical analyses highlighted weak autocorrelation and localized clustering rather than broad spatial gradients. The integration of the Radiological Risk Index provided a categorical framework that translated quantitative dose metrics into qualitative risk levels, revealing moderate risk zones at FUPRE and predominantly very low to low risk zones at PTI. In light of these findings, continued monitoring of background radiation is essential, particularly at FUPRE where infrastructural expansion is ongoing. Screening of construction materials for radionuclide content should be prioritized to mitigate potential radiological enhancement. The adoption of categorical indices such as the RRI is recommended for clearer communication of radiological risks to institutional management and regulatory agencies, ensuring that technical results are translated into actionable guidance. Extending similar spatial–statistical evaluations to other institutional environments across Nigeria would strengthen regional baselines and facilitate comparative risk assessment. Ultimately, the results underscore the importance of integrating geostatistical modeling, uncertainty analysis, and categorical risk indices into environmental radiation monitoring, thereby enhancing both scientific rigor and public health relevance. Highlights 1. Background ionizing radiation was measured across two university campuses in Southern Nigeria. 2. FUPRE exhibited marginally elevated exposure rates, linked to recent construction activities. 3. PTI maintained consistently lower exposure levels, well below UNSCEAR and ICRP reference value. 4. Spatial–statistical analysis revealed weak autocorrelation and localized hotspots at FUPRE. 5. The Radiological Risk Index provided categorical risk assessment, highlighting moderate risk zones at FUPRE and very low to low risk zones at PTI. Declaration of Competing Interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
