9种介电类水分传感器对风沙土含水率的测定精度及校准模型比较  

Performance and calibration of nine dielectric moisture sensors in measuring water content of aeolian sandy soil

作　　者：曹尤淞 肖波[1,2,4] CAO Yousong;XIAO Bo(Research Center of Soil and Water Conservation and Ecological Environment,Chinese Academy of Sciences and Ministry of Education,Yangling,Shaanxi 712100,China;State Key Laboratory of Soil Erosion and Dryland Farming on the Loess Plateau,Institute of Soil and Water Conservation,Chinese Academy of Sciences and Ministry of Water Resources,Yangling,Shaanxi 712100,China;University of Chinese Academy of Sciences,Beijing 100049,China;Key Laboratory of Arable Land Conservation in North China,Ministry of Agriculture and Rural Affairs,College of Land Science and Technology,China Agricultural University,Beijing 100193,China)

机构地区：[1]中国科学院教育部水土保持与生态环境研究中心,陕西杨凌712100 [2]中国科学院水利部水土保持研究所,黄土高原土壤侵蚀与旱地农业国家重点实验室,陕西杨凌712100 [3]中国科学院大学,北京100049 [4]中国农业大学土地科学与技术学院,农业农村部华北耕地保育重点实验室,北京100193

出　　处：《西北农林科技大学学报（自然科学版）》2024年第3期123-135,145,共14页Journal of Northwest A&F University(Natural Science Edition)

基　　金：国家自然科学基金面上项目(42077010);中国科学院“西部之光”人才培养引进计划项目(2019)。

摘　　要：【目的】比较不同水分传感器对风沙土含水率的测定精度,并建立校准模型,为旱区农业水土资源的高效利用提供理论依据。【方法】以METER、Acclima和Truebners三大制造商生产的9种介电类水分传感器为对象,以典型风沙土为供试土壤,通过室内校准试验比较各种传感器对土壤含水率的测定精度,评估其在风沙土上的适用范围、准确度、精密度及其影响因素,并构建不同传感器测定风沙土含水率的校准模型,比较不同校准模型的精度。【结果】(1)默认模型下,与其他传感器相比,MAS-1和EC-5的测定精度较高,其均方根误差(RMSE)分别为0.020和0.027,平均偏差误差(MBE)分别为0.016和0.024,斜率(k值)分别为0.9433和0.9403,决定系数(R^(2))分别为0.926和0.938。(2)土壤含水率范围影响传感器的测定精度。各传感器在低含水率下的RMSE平均值比中、高含水率分别减小了40.9%和42.6%,MBE平均值减小了61.8%和59.9%,而R2平均值提高了0.7%和11.3%。其中,低含水率时EC-5和TDR-315H的精度较高,而中、高含水率下MAS-1的精度均较高。对于含水率相同的土壤,各传感器的测定结果差异较大,且含水率越高差异越大。(3)与默认模型相比,校准模型的RMSE和MBE平均减小了48.8%和72.6%,纳什系数(NSE)和R2提高了70.7%和4.5%。经模型校准后,5TE和TEROS-12测定精度的增幅最大,而TDR-315H的测定精度相对最高。此外,传感器测定的风沙土含水率默认值与校准值具有较高的拟合精度,通过模型转换可实现对默认值的二次校准。【结论】综合评估测定精度、使用寿命和售价,MAS-1、EC-5和TDR-315H可作为风沙土含水率监测的优先传感器备选,且利用风沙土的校准模型对传感器进行标定十分重要。【Objective】This study comprehensively evaluated the accuracy of various moisture sensors in measuring water content of aeolian sandy soil and established calibration models,which was a crucial step towards the efficient utilization of agricultural water and soil resources in drylands.【Method】In this study,the laboratory calibration experiments for nine dielectric moisture sensors produced by three major manufacturers(METER,Acclima and Truebners)were conducted with typical aeolian sandy soil in the northern Loess Plateau.The applicability,accuracy and precision of each sensor were thoroughly analyzed along with the factors influencing their performance.The calibration models of sensors for aeolian sandy soil were established and validated to determine and differentiate their accuracy.【Result】(1)In contrast to other sensors,MAS-1 and EC-5 had the highest accuracy level with root mean square error(RMSE)of 0.020 and 0.027,mean bias error(MBE)of 0.016 and 0.024,slope(k)of 0.9433 and 0.9403,and coefficient of determination(R^(2))of 0.926 and 0.938,respectively.(2)The accuracy of sensor was significantly affected by water content level.On average,RMSE of each sensor at low water content was 40.9%and 42.6%lower than that of medium and high water contents,MBE was 61.8%and 59.9%lower and ave-rage R 2 was 0.7%and 11.3%higher,respectively.Specifically,EC-5 and TDR-315H exhibited superior accuracy levels at low water content,and the accuracy of MAS-1 was the highest at medium and high water contents.With identical water contents,each sensor yielded measurably distinct outcomes,and the diffe-rences increased as water content increased.(3)The calibration model decreased RMSE and MBE by 48.8%and 72.6%,while concurrently increasing Nash-Sutcliffe efficiency(NSE)and R 2 by 70.7%and 4.5%,respectively.Notably,calibrated 5TE and TEROS-12 demonstrated the highest accuracy incrementation,while calibrated TDR-315H attained the highest accuracy.The relationship between default and calibration value was well-fitted via a non-l

关 键 词：土壤体积含水率 介电常数 水分传感器 农业水土资源 生态水文过程 旱区农业 

分 类 号：S152.7[农业科学—土壤学]