期刊文献+

基于MEA-BP模型的西北地区参考作物腾发量模拟 预览

Reference Evapotranspiration Simulation in Northwest China Based on MEA-BP Model
在线阅读 免费下载
收藏 分享 导出
摘要 【目的】实现较少气象资料输入下参考作物腾发量(ET0)的精准模拟。【方法】选取西北地区5个代表性气象站点1967—2016年逐日气象数据作为输入参数,以FAO-56 Penman-Monteith(PM)模型计算的ET0作为标准值,基于思维进化算法(mind evolutionary algorithm,MEA)优化BP神经网络,构建了不同气象因子输入组合的12种MEABP模型,并将模拟结果与Penman-Monteith 24(PM 24)、Hargreaves-Samani(H-S)和Irmark-Allen(I-A)3种在西北地区ET0模拟精度较高的经验模型进行了比较。【结果】MEA-BP模型能很好地反映气象因子与ET0间复杂的非线性映射关系;MEA-BP11模型(输入气温、日照时间和风速)、MEA-BP10模型(输入气温、相对湿度和风速)和MEA-BP7模型(输入气温和风速)的R2、NSE、MAE、RMSE和nRMSE范围分别为0.978 9~0.986 5、0.977 7~0.985 6、0.172 2~0.216 6 mm/d、0.229 9~0.285 9 mm/d和3.96%~8.64%,GPI排名分别为2、3、4,精度均明显高于3种经验模型;与H-S、I-A和PM 24模型有相同气象因子输入的MEA-BP1模型(仅输入气温)的R2、NSE、MAE、RMSE和nRMSE分别为0.770 6、0.644 3、0.772 8 mm/d、1.037 2 mm/d和31.48%,MEA-BP8模型(输入气温和日照时间)的分别为0.782 4、0.669 0、0.745 2 mm/d、1.004 6 mm/d和30.47%,MEA-BP12模型(输入气温、相对湿度、日照时间和风速)的分别为0.987 5、0.986 6、0.164 8 mm/d、0.222 0 mm/d和6.71%,其GPI排名分别为8、7和1,以上3种MEA-BP模型模拟精度均明显高于相同气象因子输入下H-S、I-A及PM 24模型。【结论】在中国西北地区应用MEA-BP模型可实现较少气象参数输入下ET0精准模拟,当仅输入气温时推荐使用MEA-BP1模型,当输气温和风速时推荐使用MEA-BP7模型,当输入气温、日照时间和风速时推荐使用MEA-BP11模型。 【Objective】Achieve accurate simulation of reference crop evapotranspiration (ET0) with less meteorological data input. 【Method】The daily meteorological data of 50 representatives (1967—2016) of five representative meteorological stations in Northwest China were selected as input parameters. The ET0 calculated by FAO-56 Penman-Monteith (PM) model was used as the standard value, based on the thought evolution algorithm (mind). Evolutionary algorithm (MEA) optimizes BP neural network, constructs 12 MEA-BP models with different meteorological factors input combinations, and the simulated results were compared with Penman-Monteith 24 (PM 24), Hargreaves - Samani (H-S) and Irmark-Allen (I-A), which were three kinds of empirical models with higher accuracy in ET0 in the northwestern region.【Result】The MEA-BP model could well reflect the complex nonlinear mapping relationship between meteorological factors and ET0;MEA-BP11 model (input air temperature, sunshine hours and wind speed), MEA-BP10 model (input air temperature, relative humidity and the wind speeds) and MEA-BP7 model (input air temperature and wind speed) had R2, NSE, MAE, RMSE and nRMSE ranges of 0.978 9 to 0.986 5, 0.977 7 to 0.985 6, 0.172 2 to 0.216 6 mm/d, 0.229 9 to 0.285 9 mm/d, and 3.96% to 8.64%, respectively, with GPI rankings of 2, 3, and 4 respectively. The accuracy was significantly higher than that of the three empirical models;They were the same meteorological input parameters as the H-S, I-A and PM 24 models, and the R2, NSE, MAE, RMSE and nRMSE of the MEA-BP1 model (input air temperature only) were 0.770 6, 0.644 3, 0.772 8 mm/d, 1.037 2 mm/d and 31.48%, the MEA-BPNN8 model (input air temperature and sunshine hours) were 0.782 4, 0.669 0, 0.745 2 mm/d and 1.004 6 mm/d and 30.47% respectively,the MEA-BP12 model (input air temperature, relative humidity, sunshine hours, and wind speed) were 0.987 5, 0.986 6, 0.164 8 mm/d , 0.222 0 mm/d and 6.71% respectively, whose GPI rankings were 8, 7, and 1, respectively, and the simulation acc
作者 余婷 崔宁博 张青雯 冯禹 龚道枝 胡笑涛 YU Ting;CUI Ningbo;ZHANG Qingwen;FENG Yu;GONG Daozhi;HU Xiaotao(State Key Laboratory of Hydraulics and Mountain River Engineering & College ofWater Resources and HydroPower, Sichuan University, Chengdu 610065, China;Provincial Key Laboratory ofWater-saving Agriculture in Hill Area of Southern China,Chengdu 610066, China;State Engineering Laboratory for EfficientWater Use and Disaster Loss Reduction of Crops, Institute of Environment and Sustainable Development in Agriculture, Chinese Academy of Agriculture Science, Beijing 100081, China;Institute ofWater-saving Agriculture in Arid Areas of China, Northwest A&F University, Yangling 712100, China)
出处 《灌溉排水学报》 CSCD 北大核心 2018年第S2期107-115,共9页 Journal of Irrigation and Drainage
基金 “十三五”国家重点研发计划课题(2016YFC0400206) 国家自然科学基金项目(51779161) “十二五”国家科技支撑计划课题(2015BAD24B01) 2017 年中央高校基金科研业务费专项(2016CDDY-S04-SCU,2017XDLZ-N22).
关键词 参考作物腾发量 MEA-BP神经网络 中国西北地区 模型模拟 reference crop evapotranspiration MEA-BP neural network Northwest China model simulation
作者简介 余婷(1993-),女,陕西麟游人。硕士研究生,主要从事节水灌溉理论与技术研究。E-mail:yuting_SCU@163.com;通信作者:崔宁博(1981-),男,陕西凤翔人。副教授,博士,主要从事节水灌溉理论与技术研究。E-mail:cuiningbo@126.com
  • 相关文献
投稿分析

相关作者

内容加载中请稍等...

相关机构

内容加载中请稍等...

相关主题

内容加载中请稍等...

浏览历史

内容加载中请稍等...
;
使用帮助 返回顶部 意见反馈