土壤上方快中子强度与土壤水分成反比，宇宙射线快中子技术通过记录仪器周边快中子强度信息来反演土壤水分数据。该技术是一种新型、无损的土壤水分测量方式，能够提供实时的长时间序列的土壤水分数据。本论文主要研究宇宙射线快中子技术在农田尺度土壤水分观测中的应用及其影响因素。土壤水分在陆面过程、气候过程、生态过程以及作物生长过程等模型中发挥着重要作用。传统的土壤水分测量方法多为点测量，其空间代表尺度比较小，在获取较大尺度的土壤水分信息时，工作量大，费时费力，难以在短时间内完成。相比点测量，微波遥感测量土壤水分具有时空涵盖广的优势，但是其测量的深度仅限于土壤表层。因此，在传统点测量和微波遥感大尺度测量之间存在一个缺口，即无法获取农田或小流域尺度上的区域土壤水分信息。同样，在验证遥感反演的土壤水分过程中无法获取所需的像元尺度土壤水分数据。已有的研究工作表明，探地雷达面反射法和地面波法能够提供面状土壤水分信息，但其测量周期较长，在农田下垫面下难以开展工作。近年来发展的宇宙射线快中子技术很好的解决了这些问题，该技术能探测半径三百米范围内的土壤水分信息，其有效代表深度可达几十厘米，在中尺度或农田尺度土壤水分探测中有较大的开发潜力，可为遥感反演土壤水分提供区域地面实测真值。针对上述内容，本文主要内容包括：1、 在以精准农业或节水农业为需求的土壤水分快速测量中，介绍基于宇宙射线快中子技术的土壤水分测量方法原理。2、 在理论研究的基础上开展野外观测试验，作为2012年黑河试验的一部分，在甘肃省张掖市大满灌区农田中架设宇宙射线土壤水分观测系统仪器，采集快中子数据，并采集仪器源区范围内的土样数据等。3、 对原始宇宙射线快中子数据进行质量控制，并进行气压订正，根据试验期内获取的土壤样本数据对仪器进行率定，最后计算得到2012年6月1日-9月30日作物生长季内、连续、长时间序列的土壤水分信息。结合物候数据分析作物不同生长阶段源区土壤水分的变化特征，其中灌溉和降水是影响源区土壤水分变化的主要因子。4、 宇宙射线土壤水分在灌溉期呈现一定的规律性变化，结合源区灌溉数据研究发现灌溉期宇宙射线土壤水水分呈明显的双峰变化，其变化原因与源区内三个不同社区农田的灌溉顺序有关。5、 宇宙射线快中子技术测量的土壤水分数据是一种面状数据，与单站观测的土壤水分数据有所不同，结合源区内大满超级站土壤水分数据和源区内SoilNET土壤水分数据对比发现，宇宙射线土壤水分在灌溉期呈双峰变化趋势，经历一段时间后其土壤水分达到最大值，而单点土壤水分在灌溉期为单峰变化，且在灌溉发生时土壤水分立即升高。6、 宇宙射线快中子仪器源区内布设有19个SoilNET土壤水分/温度探测测点，根据源区内单站测量的土壤水分数据采用算术平均得到源区平均土壤水分，验证宇宙射线土壤水分观测的可靠性。验证发现，在灌溉期受宇宙射线测量原理的影响，宇宙射线土壤水分与SoilNET测点平均土壤水分的均方根误差较大，在剔除灌溉期数据后，两者的相关性比较好，均方根误差由0.0372m³/m³降低到0.0275m³/m³。7、 对宇宙射线源区不同距离处的影响因子进行了分析，研究认为影响因子与距离存在反比例关系，对源区内20个测点土壤水分使用e指数倒数加权方式聚合得到的源区土壤水分更接近于宇宙射线土壤水分。

The intensity of the fast neutrons above the ground is sensitive to water content changes, largely insensitive to soil chemistry and inversely correlated with hydrogen content of the soil. By this passive, non-invasive and intermediate scale measurement, soil moisture at a horizontal scale of around 660 m and depths of 12 to 76 cm can be inferred. In this thesis, the main work are about the basic theory of cosmic-ray fast neutron probe and its agricultural application in field soil moisture measurement. Soil water content is one of the parameters to represent the land surface process, the climate process, the ecology process as well as the crops growth process. Soil water content is very heterogeneous, and conventional measurements techniques provide only ‘point’ values at a limited number of locations and can not adequately characterize the water content over large areas. Additionally, most conventional measurement techniques are invasive and time-consuming to collect, and so are not well suited for characterizing the water content in areas such as contaminated sites or for time-sensitive applications. Remote sensing with either passive microwave radiometry or active radar instruments is the most promising technique for measuring soil water content variations over large regions. Because remote sensing approaches estimate water content in the uppermost 0.05m of the soil and require that the vegetation cover is minimal, remote sensing are not applicable in all types of studies. Clearly, there is a scale gap between remote sensing and conventional measurements of soil water content. At intermediate spatial scales, such as agricultural land, reliance on sparse conventional measurements or coarse remote measurements might not provide accurate soil water content information require at these scales. Though the ground penetrating radar has been conceived as the natural intermediate-scale for soil water measurements, it is time consuming when we need to measure a large filed and in long time series monitoring. As a new soil moisture measurement, cosmic-ray fast neutron technology provides a good solution to these problems. Its large footprint makes it have great development potential in mesoscale or field scale soil moisture measurement and it is necessary to make a further research in this new technology.According to these above, the main contents include:1. Introduce the basic theory of cosmic-ray fast neutron technology in soil moisture measurement. 2. It was used in the Heihe Watershed Allied Telemetry Experimental Research (HiWATER), 2012. We placed one instrument in Daman Irrigation District of Zhangye, Gansu Provice, to record the fast neuron data and collected many soil samples in the footprint of this instrument. 3. After quality control, pressure correction and instrument calibration of the original cosmic-ray fast neutron data, continuous long time series of soil moisture were provided in the maize growing season from June 1 to September 30 of 2012. The variation of soil moisture in different growing stages was analyzed combined with the phenological data. The main factors that affected the variance of cosmic-ray soil moisture were irrigation and precipitation.4. There was some regular changes with the cosmic-ray soil moisture during the irrigation period. Combined with irrigation data in cosmic-ray probe footprint, a bimodal change was showed in the cosmic-ray soil moisture figure during irrigation period. This was because that the order of irrigation of the three communities was differenct in cosmic-ray probe footprint.5. The soil moisture of cosmic-ray fast neutron method was a regional data and was different with the soil moisture by single station observation. The difference between them was analyzed in this paper by combining with soil moisture data form Daman Station and SoilNET soil moisture/temperature probes distributed in the footpint of cosmic-ray instrument. Cosmic-ray soil moisture showed a bimodal change during irrigation periods and reached the maximum soil moisture after experiencing a period of time. However, single point soil moisture was single peak change and its maximum soil moisture occurred immediately during irrigation periods6. There were 19 SoilNET probes in the footpirnt of cosmic-ray instrument. The average soil moisture of SoilNET probes was calculated with arithmetic mean. We found that root-mean-square error between cosmic-ray soil moisture and SoilNET average soil moisture was very large during irrigatin periods because of the impact of cosmic-ray measurement theory. While root-mean-square error would decrease from 0.0372 m³/m³to 0.0275m³/m³after we eliminate the data in irrigation period. 7. The impact factors of different distances region on the cosmic-ray probe were analyzed and an inverse proportion was found. The result of twenty point soil moisture by using e-index reciprocal weighted method was closer to cosmic-ray soil moisture.