利用遥感技术动态监测大面积农田土壤水分研究

为探讨利用遥感技术动态监测大面积农田土壤水分含量 ,排除农田植被和地形的干扰 ,1 997- 1 998年在甘肃省定西县岔口乡建立了监测样区 ,在地面实测 0～ 50 cm大面积农田土壤水分本底资料和收集 5幅 TM卫片影像资料加工处理基础上 ,利用地理信息系统建立了遥感信息 (NDVI和 RVI)与土壤含水量之间的遥感光谱相关模型 ,做出了观测区土壤水分含量分布图 ,得到了初步的大面积农田土壤水分宏观动态监测结果 ,基本实现了利用遥感技术大面积动态监测土壤含水量 ,有效地指导农业春耕生产和快速掌握土壤墒情状况。研究结果表明 ,在波长 60 0～ 1 0 50 nm光谱段 ,土壤含水率与光谱反射率之间存在显著的负相关关系 (α<0 .0 5) ;利用遥感技术建立的 TM光谱水分监测模型 ,其模型监测 0～ 2 0 cm土层含水量的精度达到 90 %以上 ,实际监测土壤水分精度达到 72 .3% ;在遥感监测2 0～ 50 cm土层土壤含水量中 ,利用遥感模型监测土壤水分精度达到 80 %以上 ,实际遥感监测精度达到 60 %左右 ,并且建立了热惯量与土壤含水量之间的相关模型 ,给出了决定系数 (R2 )分析结论。

SOIL WATER CONTENT ANALYSIS IMPROVEMENT WITH REMOTE SENSING TECHNOLOGY

The present paper aims to introduce its observation and investigation of the correlation between the soil moisture and the spectral vegetation in Dingxi County, Gansu, by means of remote sensing technology. As is known, soil water moisture observation area has been established since 1997 so as to observe and monitor the soil water content in a large scale. The observation and analysis results so far gained include spectral and TM data in 0 50 cm from 1997 to 1998 in the area. Based on the author's analysis, the paper has also prepared the soil water content distribution maps below the surface ground by means of RS and GIS. The above findings indicate that there indeed exists an obvious correlation between the soil moisture and the spectral vegetation indices of TM ( p <0.05), when the vegetation interfered with the soil moisture by RS is cast away from the mixed data, and remote sensing monitoring models of soil moisture has been made by applying the remote sensing optical method. In 0 20 cm soil, the estimating soil moisture accuracy was above 90% by the models and the actual estimating accuracy was above 72.3% from the ground. In 20 50 cm soil, the estimating soil moisture accuracy proves to be above 80% by the models while the actual estimating accuracy can reach 60% observed from the ground by the optical vegetation coverage models.