Ultra-short-term prediction of LOD using LSTM neural networks

Abstract

Earth orientation parameters (EOPs) are essential in geodesy, linking the terrestrial and celestial reference frames. Due to the time needed for data processing and combining different space geodetic techniques, EOPs of the highest quality suffer latencies from several days to several weeks. However, real-time EOPs are needed for multiple geodetic and geophysical applications. Predictions of EOPs in the ultra-short term can overcome the latency of EOP products to a certain extent. Traditionally, predictions are performed using statistical methods. With the rapid expansion of computing capacity and data volume, the application of deep learning in geodesy has become increasingly promising in recent years. In particular, the Long Short-Term Memory (LSTM) neural networks, one of the most popular Recurrent Neural Network varieties, are promising for geodetic time series prediction. In this study, we investigate the potential of using LSTM to predict daily length of day (LOD) variations up to ten days in advance, accounting for the contribution of effective angular momentum (EAM). The data are first preprocessed to obtain residuals by combining physical and statistical models. Then, we employ LSTM networks to predict the LOD residuals using both LOD and EAM residuals as input features. Our methods outperform all other state-of-the-art methods in the first eight days with an improvement of up to 43% under the first EOP Prediction Comparison Campaign conditions. In addition, we assess the performance of LOD predictions using more extended time series to consider the improvements of EOP products over the last decade. The results show that extending data volume significantly increases the performance of the methods.

Geophysically Informed Machine Learning for Improving Rapid Estimation and Short‐Term Prediction of Earth Orientation Parameters

“…Since the rapid EOPs are the basis of many prediction algorithms for EOPs (Gou et al, 2023;Kiani Shahvandi & Soja, 2022a, 2022b, the results presented in this paper can be used to further enhance the prediction of EOPs, particularly in short-term horizons.…”

“…Regarding the EAM data, both the observations and forecasts are used, since forecasts can help significantly to improve the EOP prediction performance (Gou et al, 2023;Modiri et al, 2020). Since the horizon of the forecasts is also a determining factor (Kur et al, 2022), we use 14-day forecasts of ETH Zurich (Kiani Shahvandi, Gou, et al, 2022; since they are both accurate and cover a reasonable forecasting horizon for short-term EOP prediction (i.e., suitable for accurate real-time purposes).…”

“…In order to evaluate the prediction accuracy, we use the mean absolute error (MAE) metric, which is commonly used in EOP prediction studies (Gou et al, 2023;Kalarus et al, 2010;Kur et al, 2022;Modiri et al, 2018). This is done for each day individually.…”

“…There have been successful applications of machine learning for the analysis and prediction of EOPs (Dill et al, 2021;Gou et al, 2023;Kiani Shahvandi & Soja, 2022a, 2022b. Here, however, we need to consider the specific aspects of the problem and develop a new machine learning algorithm.…”

Modeling the Differences between Ultra-Rapid and Final Orbit Products of GPS Satellites Using Machine-Learning Approaches

“…The element-wise product is represented by . In this study, all the RNN architectures were combined with the LSTM neurons since their performance on predicting geodetic time series has been demonstrated [46][47][48][49].…”