灞桥樱桃始花期预测模型研究及应用

Research and application of the prediction model for the first-flowering dates of Baqiao cherry

为了实现西安市灞桥樱桃始花期有效预报及科学应用,通过创建灞桥樱桃物候观测数据库,得出新的灞桥樱桃始花期年份日序数序列。分析结果显示,2004—2020年灞桥樱桃始花期有明显的提早现象。基于灞桥樱桃主产区周围气象资料,选取诸多有影响的气象因素,运用相关分析法筛选出冬季气温、≥5℃初日和春季气温为其建模的主要因子;采用偏最小二乘回归分析方法(PLSR)建立灞桥樱桃始花期预测模型,通过回代及预测检验,结果表明预测效果较为理想。

In order to realize the effective prediction and scientific application of Baqiao cherry initial flowering period in Xi’an,a new chronological sequence of Baqiao cherry initial flowering period was obtained by creating a phenological observation database of Baqiao cherry.The results showed that the initial flowering period of Baqiao cherry was obviously earlier in 2004—2020.Based on the meteorological data around Baqiao cherry production area,many influential meteorological factors were selected,and the winter temperature,the first day≥5℃of temperature and spring temperature were selected as the main factors for modeling by using the correlation analysis method.The prediction model of Baqiao cherry’s initial flowering period was established by PLSR.The results showed that the prediction effect was ideal.