利用GPS-IR监测土壤含水量的反演模型

雪水当量的监测对于气候变化的预测、水资源管理、农业生产规划具有重要意义。GPS干涉反射(GPS interferometric reflectometry, GPS-IR)技术是一种十分有效的地表积雪监测技术，基于GPS-IR技术提出了一种雪水当量的快速估计方法。首先基于GPS-IR技术获取美国板块边界观测（plate boundary observatory，PBO）GPS站的雪深时间序列；然后利用美国积雪遥测（SNowTELemetry, SNOTEL）站观测数据构建雪水当量转换模型；最后以北美历史与预测气候数据项目（historical and projected climate data for North America，ClimateNA）的气候预测数据作为参数约束，将GPS日雪深快速转化为雪水当量，并对雪水当量估计与验证过程的影响因素进行评价。实验结果表明，基于GPS-IR技术得到的雪深序列具有良好可靠性，与观测值的相关系数（R²）达到0.98，均方根误差（root mean square error, RMSE）为11.1 cm，偏差（Bias）为-3.7 cm；快速转化模型对雪水当量估计具有较高精度（R²=0.98，RMSE=4.2 cm，Bias=-2.5 cm）与稳定性；转化模型时空稳定性较高，残差集中在5 cm内；气候预测数据的引入、积雪分布差异对雪水当量估计与验证影响较小。所提方法在积雪监测设备缺乏区域可实现雪水当量快速估计，同时也为现有积雪观测网络增强、积雪产品改善等研究提供参考。

Snow Cover and Atmospheric Variability: Changes in the Snow Covering the Earth's Surface Affect both Daily Weather and Long-Term Climate

  Objectives  Snow water equivalent (SWE) plays a key role in climate change prediction, water resource management and agricultural production planning. GPS interferometric reflectometry (GPS-IR) has been proven to be a powerful tool to monitor snow depth. An efficient framework is presented to rapidly estimate SWE from GPS-IR derived snow depth.  Methods  Firstly, the daily snow depth product is obtained using GPS observations with GPS-IR technique. Secondly, an SWE conversion model is constructed using snow depth, SWE and climate observations from snow telemetry (SNOTEL) stations in the study region. Finally, using the climate forecast data provided by the historical and projected climate data for North America (ClimateNA) project as parameter constraints, the daily GPS snow depth product is converted into SWEs.  Results  The application of the proposed framework to GPS data from the plate boundary observatory (PBO), USA shows that GPS-IR derived snow depth product is reliable (R² = 0.98, RMSE(root mean square error)= 11.1 cm, Bias =-3.7 cm).The daily GPS snow depth can be converted into SWE with high reliability (R² = 0.98, RMSE = 4.2 cm, Bias =-2.5 cm) and stability. The rapid conversion model has high spatiotemporal stability with most of the residuals lying with the range of [-5, 5] cm. The additional uncertainties introduced by the climate forecast data and the spatial variations of snow depths have limited impacts on the estimation of SWE.  Conclusions  It is believed that the proposed framework can not only provide guidance to rapidly estimate SWE inregions lacking snow monitoring equipment, but also provide a reference to enhance existing snow observation network and improve accumulated snow products.