Abstract:

According to the current situation that the probability distribution and periodic-change law of climate factors have different differences in different growth stages of tobacco leaves in field stage, the least square method was used to study the interannual trend of climate factors in different growth stages of tobacco leaves from April to September in 2020, based on the climatic statistical data of 7 national meteorological stations in Shiyan tobacco-growing area of Hubei Province. Secondly, the kernel density estimation function was used to study the probability distribution characteristics of climate factors at different growth stages. Finally, wavelet transform algorithm was used to study the difference of multi-time scale periodic transformation. The results showed that the climatic factors in different growth stages of tobacco had certain age differences, and the temperature conditions were relatively stable. The sunshine duration showed a decreasing trend in prosperous period, and the sunshine duration in other growth period was stable. There were great differences in precipitation in different growth stages. The probability distribution of temperature and precipitation in different growth stages of tobacco leaves was different, and the probability distribution of sunshine hours was concentrated. There were multi-time scale periodic changes in climate factors at different growth stages, and there was a long time scale first main cycle change in tobacco ripening stage, which was much more sensitive to climate factors than the other stages. The research findings provide a more scientific guidance for tobacco cultivation, optimizing the layout and enhancing the quality and efficiency of tobacco production.

Key words: flue-cured tobacco, tobacco growth stages, least squares, kernel density estimation, wavelet transform, climate, climatic factors, probability distribution, cyclical variations