油菜花期物候主要限制因子分析及预报模型的构建

Analysis on Limiting Factors and Construction of Prediction Model of Oilseed Rape Flowering Phenology

以油菜开花开始时间(即始花期)为研究对象,利用1980—2016年江苏省油菜发育期观测资料和同步气象数据,分析油菜始花期的时空变化趋势。结合油菜生理特性,筛选与油菜始花期显著相关的气象因子。以高淳站为例,采用通径分析明确各气象因子对油菜始花期的作用并评估因子敏感性,最后构建油菜始花期的回归预报模型。结果表明:在气候变化背景下江苏地区油菜始花期呈逐年提前趋势,淮河以南地区平均提前3 d以上;上年日平均气温稳定通过0℃终日(X 1)、当年日平均气温稳定通过5℃初日(X 2)、现蕾至开花前时段内分别大于0℃、5℃和10℃的有效积温(X 3、X 4和X 5)、最低气温分别小于0和5℃的日数(X 6和X 7)及平均最低气温(X 8)等8个因子与油菜始花期相关性达极显著水平;其中2月上旬至3月上旬的平均最低气温(X 8)、日最低气温小于5℃日数(X 7)、大于5℃有效积温(X 4)对油菜始花期的直接影响位列8个因子的前三位,且这3个因子两两共同对油菜始花期的相对影响程度也排在各因子对回归方程R 2总贡献率的前三位;其余5个因子直接效应普遍小于间接效应,且它们主要通过X 4、X 7、X 8对油菜始花期产生影响,而X 4、X 7、X 8也通过这5个因子产生一定影响;去掉任何一个因子,都会引起其他某些因子对油菜始花期的直接、间接作用发生变化;基于这8个因子构建的回归模型可解释68.48%的油菜始花期变化,并具有区域适用性。因此,就江苏地区而言,光照和降水对油菜开花早晚影响较小,热量条件才是江苏油菜开花早晚的主要限制因子,所构建的始花期预报模型可较好反映油菜开花时间早晚与相关热量因子变化的规律。

Taking the beginning dates of flowering(BDF)as the research object,this paper analyzed the spatio-temporal variation trend of the BDF in Jiangsu Province from 1980 to 2016 by using the observation data and synchronous meteorological data.Combined with physiological characteristics of oilseed rape,the meteorological factors significantly related to the initial flowering period were selected,and the effect and sensitivity of meteorological factors were defined by path analysis.Moreover,the regression prediction model of initial flowering period was constructed.The results showed that under the background of climate change,the BDF shows a tendency of advancing year by year,with an average advance of more than three days in regions south of the Huaihe River.There are eight factors,that are significantly positively correlated to the BDF,which include:the beginning date of the daily average temperature(DAT)above 0℃stably in previous year(X 1),the ending date of the DAT above 5℃stably(X 2),the effective accumulated temperature above 0,5 and 10℃(X 3,X 4 and X 5),the number of days with daily minimum temperature below 0 or 5℃(X 6 and X 7)and average minimum temperature(X 8)between budding stage and flowering stage(from early February to early March)in current year.The directly effects of X 8,X 7 and X 4 rank the top three,and the relative determination degree of these three factors to the initial flowering period of rape is also ranked among the top three factors to the total contribution rate(TCR)of all factors to the R 2 of the regression prediction model,while the direct effect of the remaining five factors are generally weaker than the indirect effect.And the remaining five factors mainly affect the BDF through X 4,X 7,X 8,while X 4,X 7,X 8 also exert some influence through these 5 factors.Removing any factor would change the direct and indirect effects on the BDF.In addition,the regression prediction model constructed by the above eight factors can explain 68.48%of the changes of the BDF(e.g.Gaochun Region),which is also suitable for some other regions.As far as Jiangsu Province is concerned,light and precipitation have little influence on the BDF,while heat condition is the main limiting factor.In a word,the prediction model of the BDF constructed in this paper can better reflect the rule of the BDF and the change of related heat factors.