Prediction of Tropospheric Ducting for UHF Signals In Plateau State Using Adaptive Neuro-Fuzzy Inference Systems (ANFIS)

Abstract

A study on the occurrence of tropospheric ducting for UHF signal in Plateau State using ANFIS had been investigated using the variation of the meteorological data; temperature, pressure, and relative humidity collected from NIMET for Jos, Plateau State Nigeria. The refractivity gradient was calculated using MATLAB as computational software. The conditions of occurrence of path clearance and tropospheric ducting were utilized to checkmate the occurrence of ducting over the period. The data gotten is used to train an ANFIS model, taking 70% of the data for model training, 20% for testing and 10% for validation. The integrity of the ANFIS model was tested using root mean square error (RMSE) and maximum absolute percentage error (MAPE) which confirmed that the model is accurate to 99% considered fit for forecasting. The developed model was able to mimic the tropospheric ducting trend up to a period of 7 years (2019- 2025) after successful simulation with an assertion that the rainy months are more prevailed to chances of tropospheric ducting occurrence as higher humidity prevails during this time which has more significance on refractivity gradient. An investigation was carried out on the effect of variation of tropospheric parameters; temperature, pressure and humidity on refractivity gradient using linear regression and Spearman correlation to ascertained the degree of dependence of each parameter and the result showed that refractivity gradient is strongly influenced by relative humidity with a correlation coefficient of -0.9772 as against temperature and pressure having correlation coefficients of -0.2587 and 0.2866 respectively.