| 摘要: |
| 目的 利用光学遥感数据获取的植被供水指数来反演西藏那曲地区的土壤湿度，结合高分辨率的遥感数据（GF-1）和中低分辨率的遥感数据（Landsat、MODIS）分别建立土壤湿度反演模型，通过比较不同空间尺度反演模型的精度和适用性，拓宽国产高分遥感数据在农牧业信息定量获取等方面的应用范围，为“天地网一体化”的现代农业信息获取和农情信息遥感监测提供理论基础。方法 以西藏那曲地区为研究区，以代表高、中、低分辨率卫星数据的高分一号（GF-1）、Landsat-8及MODIS影像数据和土壤湿度实测数据为数据源，利用植被供水指数（Vegetation Supply Water Index，VSWI）构建土壤湿度反演模型，比较3种遥感影像在反演土壤湿度方面的差异。结果 （1）VSWI反演土壤湿度的最佳深度为10 cm左右；（2）基于GF-1、Landsat-8和MODIS构建的反演模型得到的土壤湿度预测值与实测值的均方根误差分别为5.145、5.227和6.298，可见GF-1和Landsat-8的反演效果相当，均优于MODIS的反演效果；GF-1土壤反演模型的拟合效果最佳；（3）研究区土壤湿度在空间上呈东南向西北递减的趋势，与实地采样点的土壤湿度分布趋势一致，说明利用高分辨率遥感数据监测土壤湿度是可行的。结论 利用GF-1遥感数据和植被供水指数可以实现对藏北地区的土壤湿度反演，研究结果可以为干旱或者半干旱地区大范围的土壤墒情监测提供理论依据和实践参考。 |
| 关键词: 土壤湿度 反演 植被供水指数 藏北地区 |
| DOI：10.12105/j.issn.1672-0423.20180408 |
| 分类号: |
| 基金项目:国家自然科学基金项目（41401232）、农业部农业信息技术重点实验室开放课题（2016002）和华中师范大学中央高校基本科研业务费CCNU18TS002国家自然科学基金项目（41401232）、农业部农业信息技术重点实验室开放课题（2016002）和华中师范大学中央高校基本科研业务费（CCNU18TS002） |
|
| Soil moisture inversion in the Northern Tibet based on vegetation supply water index |
|
Zhou Xiaofeng1,Nie Yan1,Liu Xiuyun1,Liang Meiying1,Tan Ying1,Yu Jing2
|
|
1.Hubei Provincial Key Laboratory for Geographical Process Analysis and Simulation，Central China Normal University，Wuhan 430062，China;2.College of Resources and Environment，Hubei University，Wuhan 430062，China
|
| Abstract: |
| Purpose In this paper，the VSWI （Vegetation Supply Water Index） is obtained from different sources of optical remote sensing images and is used to invert soil moisture. Moreover，the accuracy is compared among different data sources，in order to assess their applicability and to provide implications for both theory and application development in agricultural remote sensing.Methods By using GF-1，Landsat-8 and MODIS as the data sources representing of satellite data with high，medium and low resolution，we choose Naqu County in Tibet as the study area，three soil moisture inversion models were built based on VSWI to compare different remote sensor’s ability on soil moisture inversion and analyze the uncertainty of the models.Result The results showed as follows：firstly，the optimal depth of the inversion model for soil moisture based on VSWI is 10 cm. Secondly，the model based on GF-1 data performs better than the other models in the soil moisture inversion. The root mean square errors between the field-surveyed soil moisture and the satellite-derived soil moisture are 5.145，5.227 and 6.298 for the GF-1，Landsat-8 and MODIS respectively. GF-1 and landsat-8 data have shown a similar inversion effect，which are both better than MODIS. It indicates that medium and high resolution remote sensing data has the advantages in soil moisture inversion. According to the results of inversion，the soil moisture in the study area is declining from southeast to northwest，which is consistent with the in-situ soil moisture monitoring results.Conclusion This research indicates that VSWI is an ideal indicator for estimating soil moisture and is able to provide theoretical basis and practical reference for future works，such as the larger scale inversion of soil moisture in the northern Tibet region. |
| Key words: soil moisture inversion Vegetation Supply Water Index Northern Tibet |
