摘要
为了研究豫北地区气候变化对冬小麦播种面积和产量的影响,利用Google Earth Engine(GEE)云平台对1985—2019年Landsat影像数据进行特征构建,采用随机森林(Random forest,RF)算法提取冬小麦分布信息,分析豫北地区冬小麦播种面积的变化,再结合豫北地区已知的9个气象站气象观测数据,分析播种期、拔节-抽穗期和抽穗-乳熟期内的平均气温、降水、日照时数时空变化趋势,定量统计研究区播种面积和产量的数据,对气候因子变化与冬小麦播种面积和产量进行相关性分析、重要性评价并分析产量对气候的敏感性。结果表明,RF算法分类提取结果相对误差均在3.5%以内,样本点平均精度为94.94%,Kappa系数在0.88~0.94,达到了较高精度水平。1985—2019年冬小麦总体播种面积呈上升趋势,主要分布在中部和东部地区,且中部地区播种频率较高,播种期升温对播种面积具有促进作用(P=0.013<0.05),气候因子对播种面积的重要性排序为:播种期平均气温>播种期日照时数>播种期降水。豫北地区冬小麦整体呈增产趋势,拔节-抽穗期的降水对冬小麦产量具有显著促进作用(P=0.003<0.01)且贡献率相对较高,日照时数对产量具有显著抑制作用(P=0.022<0.05);而冬小麦产量对抽穗-乳熟期日照时数的敏感性最强。气候因子对豫北地区冬小麦产量影响的重要性排序为:拔节-抽穗期降水>拔节-抽穗期平均气温>抽穗-乳熟期降水>拔节-抽穗期日照时数>抽穗-乳熟期日照时数>抽穗-乳熟期平均气温。
In order to study the impact of climate change on the sown area and yield of winter wheat in Northern Henan,the Landsat image data from 1985 to 2019 were constructed by using Google Earth Engine(GEE)cloud platform.The distribution information of winter wheat was extracted by random forest(RF)algorithm,and the changes in the sowing area of winter wheat in northern Henan were analyzed.Combined with the meteorological observation data of 9 known meteorological stations in northern Henan,spatiotemporal trends of the average temperature,precipitation,sunshine hours at the sowing date,jointing-heading stage,and heading-milk ripening stage were analyzed.The data of sowing area and yield in the study area were quantitatively analyzed.The correlation analysis,importance evaluation and sensitivity analysis of winter wheat yield to climate were carried out.The results showed that,the relative errors of the classification and extraction results of RF algorithm were all within 3.5%,the average accuracy of sample points was 94.94%,and the Kappa coefficient was between 0.88 and 0.94,reaching a high precision level.The total sowing area of winter wheat showed an increasing trend from 1985 to2019,mainly distributed in the central and eastern regions,and the sowing frequency was higher in the central region.The temperature increasing at sowing date had a positive effect on the sowing area(P=0.013<0.05).The importance of climate factors on the sowing area was ranked as follows:average temperature>sunshine hours>precipitation at sowing date.Winter wheat yield showed a trend of increasing in northern Henan.Precipitation at jointing-heading stage significantly promoted winter wheat yield(P=0.003<0.01)with a relatively higher contribution rate,while sunshine hours significantly inhibited winter wheat yield(P=0.022<0.05).The yield of winter wheat was most sensitive to sunshine hours during heading-milk ripening stage.The order of importance of climate factors on winter wheat yield in northern Henan was as follows:precipitation at jointing-heading stage>average temperature at jointing-heading stage>precipitation at heading-milk ripening stage>sunshine hours at jointing-heading stage>sunshine hours at heading-milk ripening stage>average temperature at heading-milk ripening stage.
作者
李菲菲
汤军
高贤君
杨元维
占杨英
LI Feifei;TANG Jun;GAO Xianjun;YANG Yuanwei;ZHAN Yangying(College of Earth Sciences,Yangtze University,Wuhan 430100,China)
出处
《河南农业科学》
北大核心
2022年第8期150-165,共16页
Journal of Henan Agricultural Sciences
基金
国家自然资源部地理国情监测重点实验室开放基金资助项目(2020NGCM07)。
