A predictive model for regional zenith tropospheric delay correction

The conventional zenith tropospheric delay(ZTD)model(known as the Saastamoinen model)does not consider seasonal variations affecting the delay,giving it low accuracy and stability.This may be improved with adjustments to account for annual and semi-annual variations.This method uses ZTD data provided by the Global Geodetic Observing System to analyze seasonal variations in the bias of the Saastamoinen model in Asia,and then constructs a model with seasonal variation corrections,denoted as SSA.To overcome the dependence of the model on in-situ meteorological parameters,the SSA+GPT3 model is formed by combining the SSA and GPT3(global pressure-temperature)models.The results show that the introduction of annual and semi-annual variations can substantially improve the Saastamoinen model,yielding small and time-stable variations in bias and root mean square(RMS).In summer and autumn,the bias and RMS are noticeably smaller than those from the Saastamoinen model.In addition,the SSA model performs better in low-latitude and low-altitude areas,and bias and RMS decease with the increase of latitude or altitude.The prediction accuracy of the SSA model is also evaluated for external consistency.The results show that the accuracy of the SSA model(bias:-0.38 cm,RMS:4.43 cm)is better than that of the Saastamoinen model(bias:1.45 cm,RMS:5.16 cm).The proposed method has strong applicability and can therefore be used for predictive ZTD correction across Asia.

Improvement of the prediction accuracy of polar motion using empirical mode decomposition

Previous studies revealed that the error of pole coordinate prediction will significantly increase for a prediction period longer than 100 days, and this is mainly caused by short period oscillations. Empirical mode decomposition(EMD), which is increasingly popular and has advantages over classical wavelet decomposition, can be used to remove short period variations from observed time series of pole coordinates. A hybrid model combing EMD and extreme learning machine(ELM), where high frequency signals are removed and processed time series is then modeled and predicted, is summarized in this paper. The prediction performance of the hybrid model is compared with that of the ELM-only method created from original time series. The results show that the proposed hybrid model outperforms the pure ELM method for both short-term and long-term prediction of pole coordinates. The improvement of prediction accuracy up to 360 days in the future is found to be 24.91% and 26.79% on average in terms of mean absolute error(MAE) for the x p and y p components of pole coordinates, respectively.