The Meteorological Contributing Factors To Aerosol Variability And Modelling In Nigeria

Simple item page
dc.contributor.authorBalarabe, Mukhtar Abubakar
dc.date.accessioned2017-01-12T01:00:45Z
dc.date.available2017-01-12T01:00:45Z
dc.date.issued2016-07
dc.description.abstractA long-term aerosol study is a difficult task due to limited aerosol observation stations and seasons (resulting in missing data and cloud contamination). To address these problems, this study first aimed at analyzing the characteristics and type of aerosols using long-term (1998-2013) record of aerosol optical depth (AOD) and angstrom exponent α, from ground-based Aerosol Robotic Network (AERONET). The study showed that Nigeria atmosphere is highly polluted containing mostly coarse particles as indicated by high frequency of occurrence of angstrom exponent below 1 (78 and 81%) during Harmattan (November-March) and summer (April-October). Further analysis revealed that these particles are mostly dust aerosol (DA) for both Harmattan and summer seasons (82% and 79%). Secondly, the temporal and spatial variability of the monthly mean aerosol index (AI) (qualitative indicator of the presence of dust and smoke aerosols) and meteorological parameter (wind speed, visibility, temperature and relative humidity) obtained from the Total Ozone Mapping Spectrometer (TOMS) and Ozone Monitoring Instrument (OMI) during the period of 1984-2013 for Nigeria were analyzed. The meteorological data were obtained from the National Oceanic and Atmospheric administration-National climate data center (NOAA-NCDC). The results show that the monthly mean AI has a distinct annual cycle in each zone of Nigeria (Sahel, North central, Southern and Coastal), with lowest values during the summer season and the highest values during the Harmattan season. It also revealed a significant increasing trend of AI with Corresponding decreasing trends of visibility for every zone and season. An increasing trend in annual and seasonal temperature, wind speed and relative humidity were also observed. Spatial analysis showed a strong seasonal pattern of the monthly distribution and variability of absorbing aerosols and meteorological parameters along a north to south gradient. Finally, a modified statistical models based on multiple linear regressions (MLR) and artificial neural networks (ANN) were developed to allow the estimation of the values of AI in Nigeria based on the data from ground observations. Available literatures showed that these are the first statistical (MLR) correlation and computer generated (ANN) models for AI prediction. The generated coefficients from the models were applied to another data set for cross-validation. The AI values for the missing years were retrieved, using TOMS monthly models, plotted and compared with the measured monthly AI cycle. The accuracies of the models were determined using the coefficient of determination R2, the root mean square error (RMSE) for calibrations and validations and the weighted mean absolute percentage error (wMAPE) calculated at the 95% confidence level. The results revealed that in each month, MLR can predict AI with high level of accuracies. Furthermore, the MLR models were also found effective for AI retrieval in the overall and seasonal data except in the TOM’s overall data and OMI summer data for southern and coastal zones. ANN showed high level of accuracies for AI prediction in both overall and seasonal data in all the zones. Comparison of MLR and ANN models revealed that ANN produced better prediction results compared to the MLR model in all the zones and seasons. Furthermore, feed forward network was found to outperform the cascade network. Therefore, the proposed models can be use for effective weather monitoring in Nigeria.en_US
dc.identifier.urihttp://hdl.handle.net/123456789/3408
dc.subjectAnalyzing the characteristics and type of aerosols usingen_US
dc.subjectlong-term (1998-2013) record of aerosol optical depth.en_US
dc.titleThe Meteorological Contributing Factors To Aerosol Variability And Modelling In Nigeriaen_US
dc.typeThesisen_US
Files
License bundle
Now showing 1 - 1 of 1
Thumbnail Image
Name:
license.txt
Size:
1.71 KB
Format:
Item-specific license agreed upon to submission
Description:
Download
Collections
Pusat Pengajian Sains Fizik - Tesis