涝渍胁迫条件下水分生产函数研究  

Water Production Function Under Waterlogging Stress

作　　者：梅海鹏 董国强 李伟 汪智群 周超 MEI Hai-peng;DONG Guo-qiang;LI Wei;WANG Zhi-qun;ZHOU Chao(Anhui and Huaihe River Institute of Hydraulic Research(Anhui Provincial Water Conservancy Engineering Quality Testing Center Station),Bengbu 233000,Anhui Province,China;Key Laboratory of Water Conservancy and Water Resources of Anhui Province,Bengbu 233000,Anhui Province,China;Anhui Huaihe Water Technolgy co.,ltd.Bengbu 233000,Anhui Province,China)

机构地区：[1]安徽省(水利部淮河水利委员会)水利科学研究院(安徽省水利工程质量检测中心站),安徽蚌埠233000 [2]水利水资源安徽省重点实验室,安徽蚌埠233000 [3]安徽淮河水资源科技有限公司,安徽蚌埠233000

出　　处：《节水灌溉》2024年第8期68-75,84,共9页Water Saving Irrigation

基　　金：国家自然科学基金青年项目(52109048);水利部重大科技项目(SKS-2022066)。

摘　　要：为预估涝渍综合胁迫对农业生产的影响,以皖北平原典型作物夏大豆和典型涝渍胁迫过程先涝后渍为研究对象,依据2021年6-10月在五道沟水文水资源实验站作物与水关系试验场开展的夏大豆涝渍胁迫试验,建立了涝渍胁迫程度与作物产量间的关系。通过设置不同的涝渍胁迫形式及程度的试验,分析单涝、单渍的理论减产作用,对涝渍综合胁迫过程中的涝害指标SFW和渍害指标SEW_(30)进行统一,对多个水分生产函数进行改进,并对函数参数进行率定、验证及评价。结果表明:在涝渍胁迫敏感因子识别方面,由改进的Blank模型、Singh模型和Hiler模型得到的作物对涝渍胁迫敏感程度更为可靠。在涝渍胁迫条件下的产量预测方面,以Blank模型对大豆不同生育期受涝渍综合胁迫时产量预测精度最高,综合绝对平均误差(MAE)为0.116,并且纳什系数(NSE)最接近1,模型质量最好,其次为Jensen模型、Hiler模型和Singh模型;Stewart模型质量最差,对涝渍胁迫下大豆产量预测MAE达到0.294,RMSE达到0.323,不适合作为涝渍胁迫下水分生产模型使用。研究内容为扩展涝渍胁迫条件下作物水分生产函数计算方法提供了参考,为涝渍综合胁迫下作物相对产量预测提供了理论依据。To predict the impact of integrated waterlogging stress on agricultural production,we conducted a study focusing on a typical crop,summer soybean,in the Northern Plain of Anhui Province.Our research specifically examined the waterlogging stress process involving waterlogging followed by submergence.We perform a waterlogging stress test of summer soybean performed at the crop-water relationship test site of Wudaogou Hydrological Experimental Station from June to October 2021 and describe the relationship between waterlogging stress and crop yield.Through experiments with different forms and degrees of waterlogging stress,the theoretical yield reduction effects of surface waterlogging and underground waterlogging were analyzed,and the surface waterlogging index SFW and the underground waterlogging index SEW_(30) in the process of comprehensive waterlogging stress were unified.We have improved water production function and finished the calibration,verification and evaluation of the function parameters.Results showed that in the identification of waterlogging stress-sensitive factors,the crop sensitivity to waterlogging stress obtained by the improved Blank model,Singh model and Hiler model was more reliable.In terms of yield prediction under waterlogging stress,the Blank model was used to predict the yield of soybean under comprehensive waterlogging stress at different growth stages with the highest accuracy.The comprehensive mean absolute error(MAE)was 0.116,the Nash-Sutcliffe efficiency coefficient(NSE)was closest to 1,indicating that the quality of the proposed model was the best,followed by that of the Jensen model,Hiler model and Singh model.The Stewart model has the worst quality,the predicted MAE of soybean yield under waterlogging stress reaches 0.294,and RMSE reaches 0.323,which is not suitable for use as a water production model under waterlogging stress.The research content of this paper provides a reference for expanding the calculation method of the crop water production function under waterlogging stre

关 键 词：夏大豆 涝渍胁迫 水分生产函数 先涝后渍 

分 类 号：S274[农业科学—农业水土工程] S152.7[农业科学—农业工程]