关键词: 油菜, 始花期, 花期预报, 格兰杰因果关系检验

Abstract: Rape flower festival has become a special ecological agricultural tourism project in many places of China. Accurate prediction of flowering dates can guide the public to arrange their travel time reasonably. Based on the observation data of Brassica napus  growth period from the Agricultural Meteorological Station of Menyuan, Qinghai during 1980-2021, combined with the meteorological data and the atmospheric circulation index data in the Northern Hemisphere, we analyzed the influencing factors of first-flowering date of Brassica napus in Menyuan and established the prediction model of first-flowering date using the correlation analysis and Granger causality test. The results showed that the first-flowering date of Brassica napus in Menyuan showed substantial fluctuation in advance and small fluctuation in delay. The overall trend was not obvious and the variation rate was small. The average first-flowering date was July 3 to 4. The first-flowering date was significantly correlated with average temperature, maximum temperature, accumulated temperature of ≥0 ℃, accumulated temperature of ≥3 ℃ and accumulated temperature of ≥5 ℃ (P<0.01), as well as Northern Hemisphere Subtropical High Northern Boundary Position Index, Pacific Subtropical High Northern Boundary Position Index, Pacific Polar Vortex Area Index and the East Asian Trough Position Index (P<0.05). Northern Hemisphere Subtropical High Northern Boundary Position Index, Pacific Subtropical High Northern Boundary Position Index, Pacific Polar Vortex Area Index, and the East Asian Trough Position Index had a statistically causal relationship with the initial flowering period of Brassica napus in Menyuan, with a lag of one year. Using two atmospheric circulation indices coupled with accumulated temperature of one-year lag period as the forecasting factors, the prediction equation of initial flowering date of Brassica napus in Menyuan was constructed by linear regression and principal component analysis method. This method can effectively improve the prediction accuracy within three days.

Key words: Brassica napus L., first-flowering date, flowering date prediction, Granger causality test.