两种精化的对流层延迟改正模型

对流层延迟是全球导航卫星系统(Global Navigation Satellite System,GNSS)导航定位中的重要误差源,其量值主要受气象条件影响.采用传统对流层建模思路,利用GPT2模型来提供相对准确的气温、气压和相对湿度,然后利用Saastamoinen模型来计算天顶对流层延迟,由此构建了GPT2+Saas模型;采用新的对流层建模思路,直接针对天顶对流层延迟的时空特性建模,构建了与GPT2+Saas模型相匹配的GZTDS格网模型.以GGOS Atmosphere格网数据为参考,GPT2+Saas模型(Bias:0.2cm;RMS:4.2cm)和GZTDS模型(Bias:0.2cm;RMS:3.7cm)较UNB3m模型精度分别提升34%和43%.以IGS(International GNSS Service)数据为参考,GPT2+Saas(Bias:0.5cm;RMS:4.7cm)和GZTDS(Bias:-0.3cm;RMS:3.8cm)相对UNB3m模型精度分别提升10%和27%.针对GPT2+Saas模型在少数测站出现精度异常的情况进行了研究,探讨了可能的原因.在两种不同思路构建的精化对流层模型中,GZTDS模型不仅表现出更高的精度,而且在时间稳定性和地理稳定性上也表现出优越性.

Two new sophisticated models for tropospheric delay corrections

Tropospheric delay is an important error in the Global Navigation Satellites System (GNSS) positioning and navigation, and its magnitude is mainly affected by meteorological conditions. By establishing accurate empirical models for tropospheric delay, this error can be corrected without meteorological data.#br#Adopting conventional modeling strategy, we utilize the GPT2 model to provide accurate temperature, pressure and relative humidity which are then used to calculate zenith total delay by the Saastamoinen model, thus coming to the idea of "GPT2+Saas" model. Adopting the new modeling strategy, we directly model the zenith tropospheric delay in view of its spatial and temporal characteristics, and produce the GZTDS model which is an upgrade of the GZTD model. Tropospheric delay data from "GGOS Atmosphere" and international GNSS service (IGS) are used to validate the precision and stability of the models and detailed objective precision information is shown and analyzed.#br#Taking GGOS Atmosphere data as a reference, the GPT2+Saas model (Bias: 0.2 cm; RMS: 4.2 cm) and the GZTDS model (Bias: 0.2 cm; RMS: 3.7 cm) have improved the accuracy respectively by 34% and 43% relative to the UNB3m model. Using IGS data as a reference, the GPT2+Saas (Bias: 0.5 cm; RMS: 4.7 cm) and the GZTDS (Bias: -0.3 cm; RMS: 3.8 cm) have improved the accuracy respectively by 10% and 27% relative to the UNB3m model. The GZTDS model has a better accuracy than the GPT2+Saas model by 11.9% with respect to the GGOS Atmosphere data and by 19.1% with respect to the IGS data, respectively. In the two different sophisticated models, the GZTDS not only shows a higher accuracy, but also presents higher stability in time and space.#br#According to the traditional method and the new method, we establish two different tropospheric delay models with almost the same external conditions (resolution, data source and time span). By this way, we convert the comparison of two methods into the comparison of two models. Validated by the GGOS Atmosphere data and IGS data, the new method performs better than the traditional one though both the methods are good enough, which indicates that the new method does not depend on meteorology data and can achieve as good results as the traditional ones that need meteorology data.