Abstract:

To improve the accuracy of forecasts of the full-bloom period of apricot blossoms in the Hami Oasis, which is the core of the best viewing period, and to provide a scientific basis for tourism management and visitor planning, we used 32-year apricot phenological records (1991-2022) together with concurrent surface meteorological observations. We analyzed the interannual variation in flowering dates, identified key meteorological drivers, and constructed and tested a forecasting model using principal component analysis. The results show that: (1) phenological characteristics: the mean first-flowering day-of-year (DOY) of apricot blossoms in the Hami Oasis is 92.4, corresponding to 2-3 April in common years and 1-2 April in leap years, with a range of 21 days between the earliest and latest first-flowering dates. The mean full-bloom DOY is 94.5, corresponding to 4-5 April in common years and 3-4 April in leap years, with a range of 22 days between the earliest and latest full-bloom dates. Both first-flowering and full-bloom dates exhibit a decadal advancing trend, with climatic tendency rates of -2.73 d per 10 years (r=-0.476, P<0.05) and -2.47 d per 10 years (r =-0.421, P<0.05), respectively. The temporal distribution of early versus late full-bloom dates shows a clear decadal pattern, with more early years occurring in the 21st century and more late years concentrating in the 1990s. (2) Meteorological controls: meteorological factors exert a significant influence on the optimal viewing period of apricot blossoms. The mean maximum air temperature in mid-January and early March; the mean, mean maximum and mean minimum air temperatures in mid-to-late March and for March as a whole; the mean ground temperature in mid-to-late March and for March as a whole; as well as sunshine duration and ≥5℃ effective accumulated temperature in early March all show significant or highly significant negative correlations with the full-bloom date. In other words, higher values of these variables are associated with earlier full bloom. In contrast, mean relative humidity in March; precipitation in late January and in March; and the onset dates of ≥0℃, ≥3℃ and ≥5℃ temperatures exhibit significant or highly significant positive correlations with the full-bloom date, such that higher values or later onset dates correspond to later full bloom. (3) Model performance: based on 29 years of observations from 1991 to 2019, we developed a principal component analysis forecasting model (Y=94.828-4.634x, R²=0.680). The model was validated using data of 2020-2022, yielding satisfactory performance: the forecast accuracy for full-bloom dates within 0-2 days of the observed dates reached 62.07%, and the accuracy for a 3-day difference was 10.34%. This model can provide technical support for meteorological services targeting the optimal viewing period of apricot blossoms in the Hami Oasis.

Key words: apricot flowers in Hami Oasis, blooming period, affecting factors, forecasting model