Atmospheric discharges offer great risks to the population and activities that involve different
systems such as telecommunications, energy distribution and transportation and among others.
Lightning prediction can contribute to minimize the risks of this natural phenomenon.
Therefore, this thesis presents a model for lightning prediction based on satellite atmospheric
sounding data, validated with lightning data for study areas of the Amazon region in Brazil,
through an investigation that considered five period cases for validation of lightning prediction:
case 1 (one hour), case 2 (two hours), case 3 (three hours), case 4 (four hours) and case 5 (five
hours). Two different forecasting methodologies were used: the first version of the predictor
used data from all study areas in the random formation of the sets training, validation and test.
In a second version, we did not use the criterion of randomness of the data in the formation of
the training and test sets, and same were limited for each area of the study, in order to create
individualized forecasts by geographical area studied. The machine learning technique used to
predict lightning was the Artificial Neural Network (ANN) trained with Levenberg-Marquardt
backpropagation algorithm to classify modeling related to lightning prediction. This
classification relied on the possibility of lightning prediction from the vertical profile of air
temperature obtained from satellite NOAA-19. The results obtained by RNA, in the first
approach, were compared with traditional methodologies established in the lightning prediction
literature, in the second approach the results obtained showed the predictor's output for real test
data. Results show that ANN was capable of identifying adequately the class to which a new
event belongs to in relation to categories of occurrence and absence of lightning. For the first
approach, the best performance for case 5 was obtained, with a test accuracy of 95.6%, while
for the second approach a general test accuracy of 82.04% was obtained.
|
