Semi-empirical characteristics of modified lognormal DSD inputs using rain rate distributions for radio links over the African continent

Rainfall attenuation mitigation is one of the important design factors considered in the deployment of outdoor wireless 5G networks, especially in the provision of services operating in frequency bands between 10 and 100 GHz. In this current work, previous rainfall measurement and modelling campaigns undertaken at different locations in Africa are investigated using their available rainfall Drop Size Distribution (DSD) profiles from Joss-Waldvogel (JW) disdrometer. These locations of interest are found in Western, Central and Southern Africa at the following sites: Niamey (13°30ʹN, 2°10ʹE), Boyélé (2°50ʹN, 18°04ʹE), Abidjan (5°25ʹN, 4°W), Dakar (14°34ʹN, 17°29ʹW), Ile-Ife (7°30ʹN, 18°04ʹE), Durban (29°52ʹS, 30°58ʹE) and Butare (2°36ʹS, 29°44ʹE). A variant of the lognormal DSD, categorized as a lognormal Type II DSD—is considered as a statistical tool to analyze and compare the behaviour of DSDs at these locations, using its first and second derivative conditions. Furthermore, a semi-empirical approach is proposed to formulate the input parameters of the lognormal Type II distribution using two renowned parameters: percentage of rainfall rate exceeded (P) and rainfall rate at 0.01% (R_0.01). An application of power-law regression fit is then utilized to extract the coefficients of these inputs from their DSDs at different locations in Africa. A comparison of the resulting site and climate models, with previous models using self-consistent rainfall rate predictions at 20 and 100 mm/h, shows reasonable agreement. Further validation is realized using specific attenuation prediction from Mie scattering technique with computations at frequency ranges up to 100 GHz—within the microwave and millimeter wave bands. More precisely, observations from these results show that climatic and geographic characteristics largely influence the DSD profile and expected rainfall attenuation at each location.