基于KNN-TCN模型的蒸发皿蒸发量预测研究  

Prediction of Pan Evaporation Based on KNN⁃TCN Model

作　　者：谢育珽 郑翔天 史俊才 刘萍 申文明 程文飞 李新华 杨静 邢云飞 XIE Yuting;ZHENG Xiangtian;SHI Juncai;LIU Ping;SHEN Wenming;CHENG Wenfei;LI Xinhua;YANG Jing;XING Yunfei(College of Big Data,Taiyuan University of Technology,Taiyuan 030024,China;Business School of Northeast Normal University,Changchun 130117,China;Yangquan Branch,China Mobile Communications Group Shanxi Co.,Ltd.,Yangquan 045000,China;Ministry of Ecology and Environment Center for Satellite Application on Ecology and Environment,Beijing 100094,China;College of Software,Taiyuan University of Technology,Taiyuan 030024,China;Lyuvliang Ecological Environment Bureau,Lyuliang 033000,China)

机构地区：[1]太原理工大学大数据学院,山西太原030024 [2]东北师范大学经济与管理学院,吉林长春130117 [3]中国移动通信集团山西有限公司阳泉分公司,山西阳泉045000 [4]生态环境部卫星环境应用中心,北京100094 [5]太原理工大学软件学院,山西太原030024 [6]吕梁市生态环境局,山西吕梁033000

出　　处：《人民黄河》2024年第6期113-118,125,共7页Yellow River

基　　金：山西省重点研发计划项目(202202020101007);山西省自然科学基金青年基金项目(201901D211002)。

摘　　要：蒸发量的精确预测对合理开发利用水资源、旱涝变化趋势研究和农作物灌溉用水量的估算具有十分重要的意义。选取我国北方地区14个地面国际交换站观测的7项气象数据,以时间卷积网络(TCN)模型为基础模型,运用K-近邻(KNN)算法对蒸发皿蒸发量的空间因素进行筛选,构建KNN-TCN蒸发皿蒸发量预测模型,并利用平均绝对误差、均方根误差和判定系数3项指标对目标站点的蒸发量预测精度进行评价。结果表明:1)KNN-TCN模型预测结果明显优于LSTM模型;2)相比基础TCN模型,KNN-TCN模型预测结果的判定系数提升了2.52%,平均绝对误差、均方根误差分别降低了23.97%、13.06%。Accurate prediction of evaporation is of great significance for the rational development and utilization of water resources,the study of drought and flood trends,and the estimation of crop irrigation water consumption.In this paper,7 meteorological data observed by 14 ground international exchange stations in northern China were selected.Based on the time convolution network(TCN)model,the K-nearest neighbor(KNN)algorithm was used to screen the spatial factors of pan evaporation.The KNN-TCN pan evaporation prediction model was built and the average absolute error,root mean square error and coefficient of determination were used to evaluate the evaporation prediction ac-curacy of the target site.The results show that a)the prediction results of KNN-TCN model are significantly better than that of LSTM model.b)Compared with the basic TCN model,the determination coefficient of KNN-TCN model is increased by 2.52%,and the mean absolute error and root mean square error are reduced by 23.97%and 13.06%,respectively.

关 键 词：蒸发皿蒸发量 时间卷积网络 K-近邻算法 空间因素 

分 类 号：P426[天文地球—大气科学及气象学] TP183[自动化与计算机技术—控制理论与控制工程]