The need to study the variations of climate change in Nigeria becomes necessary at a time the universe and Nigerians, in particular are passing through challenges due to climate change as a result of emissions. The atmospheric gases have a greater transparency for incoming solar radiation, while the outgoing are trapped and re-emitted back to the Earth. This study correlated between global solar radiation and greenhouse gases over Nigeria using neural network. The results showed that positive correlations exist between solar radiations: CO2 and CH4 respectively, while exhibiting negative correlations with tropospheric ozone and water vapour. Consequently, an increase in 0.1017, 0.1350 units of CO2 and CH4, respectively could enhance the trapping and transmission of solar radiation in the atmosphere, while an increase of 1.1234 and 0.1530 units of tropospheric ozone and water vapour could cause absorption of solar radiation. The trapped energy is re-radiated back to the Earth, this warms up the atmosphere and the surface of the Earth resulting to global warming. Coefficient of determination revealed that 18%, 30%, 20%, and 29%, of the variances of solar radiation being studied is explained by the variance of the water vapour, tropospheric O3, CO2, and CH4, respectively.

The refractivity profile variation in troposphere is one of the aspects that influences long-distance terrestrial electromagnetic wave propagation and performance of communication systems. This study is aimed at calculating and estimating radio refractivity at Makurdi with tropospheric parameters of relative humidity, absolute temperature and atmospheric pressure using ITU-R and artificial neural network models. Validation results are thus, absolute temperature = 0.4313 K, relative humidity = 0.9989 %, pressure = 0.0201 (hpa) respectively. The validation of the correlation coefficient results shows that all the tropospheric parameters have effects on radio refractivity, but relative humidity has more effect which is attributed to the large quantity of moisture at the troposphere. From the estimation results, it is clear that artificial neural network has the capacity of estimating tropopheric refractivity since the estimated values has close agreement with the calculated values.

No abstract

Atmospheric pollution due to carbon dioxide emission from different fossil fuels and deforestations are considered as a great and important international challenge to the societies. This study is to investigate carbon dioxide (CO2) distributions in selected points in Nigeria using neural network. Neural network model were used to estimate daily values of carbon dioxide, study spatial temporal variations of carbon dioxide, and study the annual variations of estimated and observed carbon dioxide in Nigeria. The study areas used in this work are thirty six (36) points location over Nigeria. The data used in this work is a satellite carbon dioxide () data were obtained from Global Monitoring for Environment and Security (GMES) under the programme of Monitoring Atmospheric Composition And Climate (MACC) www.gmes-atmosphere.eu/data between 2009-2014. The neural network architecture used comprises of three main layers; an input layer, a hidden layer and an output layer. Four input data were considered which include year, day of year (DOY) representing the time, latitude and longitude. Twenty hidden neurons were employed, while the output is the desired data of carbon dioxide. The results show that the increase in trend of CO2 in dry season in every part of the country is on yearly bases. In the wet season, the concentration of CO2 in Nigeria is not as much as in the dry season case, probably due to absorption of the gas by precipitation. The continuous annual increase of CO2 distribution suggests continuous increase of the greenhouse gas in Nigeria. This reveals continuous contribution of CO2 in Nigeria. The similarity in the estimated and observed signatures reveals that neural network model performance were excellent and efficient in determination of spatial distribution of CO2, thereby proving to be useful tool in modeling the greenhouse gases. The results show that neural network model has the capacity of investigating greenhouse gases variations in Nigeria.