The wide-spread exposure to constantly evolving wireless technologies believed to pose a serious health threat. Human beings are persistently exposed to RF radiation from mobile phones and their base stations. The current study aimed at classifying and characterizing the exposure to RF radiation from the mobile phone base stations. Spatial distribution measurements were carried out in Khartoum city during two time periods, first in 2012 (pilot survey) and again during Sept. 2019-Jan. 2020, to cover a total of 282 antennas operating with GSM900, GSM1800, and UMTS2100. The tested antennas belong to three mobile communication companies namely Sudani, Zain, and MTN companies, that randomly coded into company A, company B, and company C for security purposes. Measurements were performed using frequency-selective RF analyzer at fixed distances from the antennas/ towers. Data were subjected to advanced repeated measures ANOVA, linear discriminant analysis (LDA), and spatial interpolation with ArcGIS. The averages of GSM900, GSM1800, and UMTS measurements were 0.01933 W/m 2 , 0.0067 W/m 2 , and 0.0046 W/m 2 . The high levels of power densities for each single antenna were recorded at 90 m, 110 m, 130 m, and at 150 m distances, for the majority (70%) of the measured antennas and the peak/highest values reported mainly at 110 m distance. Conversely, the discriminant loadings as part of LDA, suggested that, much of variance among measurements is attributed to measurements at 150 m, 170 m, and 190 m distances, while visual illustration of group centroids implied that, the RF signals of the different companies were measured separately which support accuracy of frequency-selective measurements. The LDA has confirmed the ANOVA results that, the overall difference between the three companies was statistically significant for UMTS, and GSM900 measurements but not significant for GSM1800 measurements. Kriging interpolation using ArcGIS provided a strong evidence of great spatial distribution of exposure across the study area, with market places and typical urban residential quarters showing highest levels of RF. Few extreme values exceeding ICNIRP limits are reported but excluded from the calculations because of an issue of normality of data that is considered a prerequisite for parametric data analysis. Existence of extreme levels of RF indicates a need for further investigation and some antennas of Company B are mounted on towers belongs to Company C, implying multi exposure. Unexpected pattern of RF levels continued to increase up to 190 m distance and possibly beyond 190 m is reported for UMTS measurements of Company C.
