Minimum threshold for the sampling rate to prevent amplitude distortions in aliasing-free GPR surveys

We study sampling-related amplitude distortions within aliasing-free GPR data sets, and compare them with other factors which can affect the recorded signal. In particular, we analyze how much the sampled peak amplitudes can change with different sampling rates, and recommend a minimum threshold for the sampling rate in order to contain the maximum peak amplitude error within acceptable limits. The selection of the sampling rate during data acquisition is commonly based on the Nyquist-Shannon theorem, which offers practical lower limits in order to avoid aliasing effects and to accurately preserve the spectral content of the original analog signal. However, we show that the Nyquist-Shannon theorem does not prevent possible amplitude distortions, and that significant and unrecoverable data loss can occur even in aliasing-free data sets. We also show that interpolation and re-sampling offer only limited solutions, since the accuracy of the reconstructed signal depends on the implemented interpolation method, while its subsequent re-sampling simply reintroduces the initial problem. Based on our analysis, we recommend using during data acquisition a sampling rate equal to at least 12 times the signal central frequency, which is higher than the commonly adopted standards, in order to limit the maximum peak amplitude error within 5%