Time‐Series Prediction Approaches to Forecasting Deformation in Sentinel‐1 InSAR Data

Time series of displacement are now routinely available from satellite InSAR and are used for flagging anomalous ground motion, but not yet forecasting. We test conventional time series forecasting methods such as SARIMA and supervised machine learning approaches such as long short‐term memory (LSTM) compared to simple function extrapolation. We focus initially on forecasting seasonal signals and begin by characterizing the time‐series using sinusoid fitting, seasonal decomposition, and autocorrelation functions. We find that the three measures are broadly comparable but identify different types of seasonal characteristic. We use this to select a set of 310 points with highly seasonal characteristics and test the three chosen forecasting methods over prediction windows of 1–9 months. The lowest overall median RMSE values are obtained for SARIMA when considering short term predictions (<1 month), whereas sinusoid extrapolation produces the lowest median RMSE values for longer predictions (>6 months). Machine learning methods (LSTM) perform less well. We then test the prediction methods on 2,000 randomly selected points with a range of seasonalities and find that simple extrapolation of a constant function performed better overall than any of the more sophisticated time series prediction methods. Comparisons between seasonality and RMSE show a small improvement in performance with increasing seasonality. This proof‐of‐concept study demonstrates the potential of time‐series prediction for InSAR data but also highlights the limitations of applying these techniques to nonperiodic signals or individual measurement points. We anticipate future developments, especially to shorter timescales, will have a broad range of potential applications, from infrastructure stability to volcanic eruptions.