Seismometer Orientation Correction via Teleseismic Receiver Function Measurements in West Africa and Adjacent Islands

To establish a robust method of determining and correcting the misorientation of seismometers with radial and tangential components of the teleseismic receiver function (RF), we assess the measurement of the amplitude of P-wave phase arrivals at zero seconds (PRFamp method) and the energy of the width of the time window around the Pwave phase arrival (PRFenergy method) using varying Gaussian factors, ranging from 0.5 to 5.0 with equivalent cutoff frequencies. To understand the efficacy and discrepancies in these methods, and their results across different geological terrains, we conduct an assessment of their performance on 26 permanent seismic stations located in West Africa and adjacent islands. We conduct several tests to examine the effect of varying Gaussian factors on the estimates of misorientation. These effects are more noticeable in results from stations for which RFs are indicative of 3D or complex structures and islandic stations. Statistical analyses reveal good correlation between the results obtained from both the methods, although results from PRFamp method proves to be more stable due to their proximity to the statistical mean and standard deviation results. In comparison with the PRFenergy results, there is a smaller error difference and uncertainty, a higher correlation coefficient, a relatively lower standard deviation, and better convergence around the mean misorientation obtained using PRFamp. From both the methods, estimates from continental stations reveal similar trends, results from coastal stations tend to be consistent, whereas some island stations show a reverse trend in sensor misorientation classification. We propose that the PRFamp yields more consistent results and is less sensitive to subsurface heterogeneities. Gaussian filtering with an alpha value (alpha) of 1.5 is recommended as a trade-off between oversimplification and complications due to subsurface structures. However, we suggest that multiple approaches be adopted to provide robust bases for the estimation and comparison of sensor orientation correction calculations.