Thunderstorm hazards and their impact in the lake Victoria Basin of Kenya

Abstract

This study characterised thunderstorm hazards in the Lake Victoria Basin of Kenya. This was done by examining the relationship between thunderstorm frequency and other weather variables, analysing the storms' spatial and temporal variation, establishing the frequency characteristics of rainfall associated with the storms and assessing lightning severity in the area. The analysis involved multiple regression, coefficient of variation, Weibull formula and the lighting hazard indices respectively.

The results indicated that each thunderstorm recording station had its own multiple regression model. The coefficient of multiple determination from each station showed that the variables explained only a small fraction of the variance in thunderstorm frequency suggesting that other uninvestigated factors could be responsible for the remaining variance. Although thunderstorm frequency was high in stations adjacent to the lake, a fact attributed to relative convective activities associated with the lake, the F-test tevealed no significant difference in the number of thunder days among the stations.

A monthly analysis of the storms' frequency showed each recording station had its own peak. The first peak was evident in Kisii during the month of March followed by Kericho in May. The third peak was observed in Kitale during the month of August. The least peak was evident in Eldoret.

The Gumbel probability plots for annual maximum daily rainfall (Rmax) from thunder days extracted fromt the stations revealed Kisumu to have the highest Rmax while Eldoret had the lowest. This analysis has implications in the design of hydraulic structures in the area.

In assessing lighting severity in the Basin using the Lighting Hazard Index (LHI), Kisii and Nyamira Districts which have recorded the worst thunderstorm tragedied were taken as case studies. Since the LHI considered only human casualties, an adjusted index, Overall Lighting Hazard Index (OLHI) was developed to accommodate other losses such as houses, livestock and trees. Both indices showed that lightning seriousness was high in areas with high population densities such as Keumbu and Nyamira Divisions. Although there was no significant difference between the two indices, the study has recommended the use of OLHI as it considers both loss of life property.