Techniques for lightning prediction: A review

Abstract

Lightning is a natural occurrence which is created through the mixture of hot and cold air in the cloud. Sudden occurrence of lightning has caused damages to many lives and properties, for this reason; there is a need to develop a system that can predict lighting occurrence for people to take necessary precaution. However, accurately predicting lightning has been a challenge among researchers, as they find it difficult to select the right approach and algorithms to use when predicting lightning. Thus, this paper presents a systematic literature review on the best techniques for lightning prediction by reviewing relevant papers that are systematically collected based on the inclusion and the exclusion criteria from four different academic databases which includes Scopus, IEEE Xplore, Science direct, and SpringerLink. The findings from the review shows that the Random Forest algorithm is mostly used for lightning prediction and has generally out performed all other algorithms that have been used in lightning prediction in remote region. Also the review finds out that there is an inverse relationship between predicting system accuracy and lead time. Another observation in the research is that numerical weather prediction predicts more accurately compare to geo satellite prediction.

Keywords

• Lightning prediction; systematic literature review; numerical weather prediction; machine learning and algorithms.