In this study we demonstrate how very-long period (VLP) volcanic seismic signals can be processed in order to obtain essential and detailed information about the seismo-volcanic source process. As an example we use the VLP signal observed on 23 March 2012 during an outgassing event at Soufrière Hills volcano, Montserrat, acquired by instruments with different natural periods. The aim of this study is to highlight the importance of retrieving the correct source time function by a complete restitution process. When ground displacement cannot be retrieved through the restitution process due to very narrow band-pass limited instrument response, we compare synthetic and observed waveforms in the velocity domain and determine the best model by generating a synthetic velocity seismogram using the band-limited seismometer characteristics. Furthermore, we show how this approach of forward modelling can reveal much more detail of the source process, since small changes in displacement are enhanced in the velocity seismogram. Using these restituted and modelled displacements we perform a moment tensor inversion combined with a grid search locating the source at 600 m depth below sea level and estimating the source volume change to be in the range of 0.6  − 1.1 × 10³ m³.

在这项研究中，我们演示了如何处理非常长的时期（VLP）的火山地震信号，以便获得有关地震火山源过程的基本而详细的信息。例如，我们使用在蒙特塞拉特SoufrièreHills火山发生的一次放气事件中观察到的VLP信号，该事件是由具有不同自然周期的仪器采集的，该事件发生在2012年3月23日。这项研究的目的是强调通过完整的恢复程序来检索正确的源时间函数的重要性。当由于非常窄的带通有限仪器响应而无法通过恢复过程获得地面位移时，我们将比较速度域中的合成波形和观测波形，并通过使用带限地震仪特性生成合成速度地震图来确定最佳模型。此外，我们展示了这种正演模拟方法如何揭示源过程的更多细节，因为速度地震图中位移的微小变化得到了增强。使用这些经过恢复和建模的位移，我们进行了矩张量反演并结合了网格搜索，将震源定位在海平面以下600 m的深度处，并将震源体积变化估计在0.6的范围内  −  1.1×10 ³  m ³。