Abstract

Accurate modeling of tropospheric delays is crucial for the global navigation satellite system (GNSS), which finds extensive applications in early warning systems of natural hazards and extreme weather forecasting. Zenith tropospheric delay (ZTD) is estimated as a random walk process with a constraint in GNSS processing. The constraint, referred to as random walk process noise (RWPN), holds significant importance in real-time ZTD estimation and exhibits geographical and temporal specificity. Presently, RWPN is treated as either a constant value or derived from a numerical weather model (NWM). To address this, our study presents a global RWPN model (GRM) by parameterizing a decade of NWM-derived RWPN data. Taking into account its spatiotemporal nature, we formulate the RWPN equation for each station by employing trigonometric, exponential, and Legendre functions. The optimum RWPN value is determined by incorporating GRM using latitude, longitude, orthometric height, and time as inputs. To validate the efficacy of GRM, we compare its performance against RWPN values derived from both JRA-55 and ERA5 datasets for the year 2020. The results indicate that the GRM-derived values exhibit enhanced accuracy in comparison with the optimal fixed RWPN values, as well as the yearly and monthly mean RWPN values. Additionally, we assess the efficacy of the GRM model in real-time ZTD estimation across 20 globally distributed GNSS stations. The results reveal an improvement exceeding 10% when compared to the results of the best fixed RWPN values. The GRM model offers an effective solution for obtaining accurate RWPN values on a global level, all while minimizing computational demands and time constraints. This notable progress significantly bolsters the precision of real-time GNSS estimates, thus facilitating their application in time-sensitive geophysical and meteorological scenarios.

Highlights

A global RWPN model is introduced for real-time GNSS tropospheric delay estimation. The model provides precise RWPN values while minimizing computation cost and time. The proposed model improves the accuracy of real-time ZTD estimation by over 10%. This model promotes GNSS in time-critical geophysical applications.

中文翻译：

---

使用新颖的全局随机游走处理噪声模型 (GRM) 进行实时 GNSS 对流层延迟估计

摘要

对流层延迟的准确建模对于全球导航卫星系统（GNSS）至关重要，该系统在自然灾害早期预警系统和极端天气预报中有着广泛的应用。天顶对流层延迟 (ZTD) 被估计为 GNSS 处理中具有约束的随机游走过程。该约束称为随机游走过程噪声 (RWPN)，在实时 ZTD 估计中具有重要意义，并表现出地理和时间特异性。目前，RWPN 被视为恒定值或源自数值天气模型 (NWM)。为了解决这个问题，我们的研究通过对十年来 NWM 衍生的 RWPN 数据进行参数化，提出了一个全局 RWPN 模型 (GRM)。考虑到其时空性质，我们利用三角函数、指数函数和勒让德函数制定了每个站点的RWPN方程。最佳 RWPN 值是通过使用纬度、经度、正高和时间作为输入合并 GRM 来确定的。为了验证 GRM 的有效性，我们将其性能与从 2020 年 JRA-55 和 ERA5 数据集得出的 RWPN 值进行比较。结果表明，与最佳固定 RWPN 值相比，GRM 得出的值表现出更高的准确性，如下所示以及年和月平均 RWPN 值。此外，我们还评估了 GRM 模型在 20 个全球分布的 GNSS 站的实时 ZTD 估计中的有效性。结果表明，与最佳固定 RWPN 值的结果相比，改进超过 10%。 GRM 模型提供了一种有效的解决方案，可在全球范围内获取准确的 RWPN 值，同时最大限度地减少计算需求和时间限制。这一显着进展显着提高了实时 GNSS 估计的精度，从而促进了其在时间敏感的地球物理和气象场景中的应用。

强调

引入全局 RWPN 模型进行实时 GNSS 对流层延迟估计。该模型提供精确的 RWPN 值，同时最大限度地减少计算成本和时间。所提出的模型将实时 ZTD 估计的精度提高了 10% 以上。该模型促进了 GNSS 在时间关键的地球物理应用中的应用。