Most of the data acquisition in ground-penetrating radar is done along fixed-offset profiles, in which velocity is known only at isolated points in the survey area, at the locations of variable offset gathers such as a common midpoint. We have constructed sparse, heavily aliased, variable offset gathers from several fixed-offset, collinear, profiles. We interpolated those gathers to produce properly sampled counterparts, thus pushing data beyond aliasing. The interpolation methodology estimated nonstationary, adaptive, filter coefficients at all trace locations, including at the missing traces’ corresponding positions, filled with zeroed traces. This is followed by an inversion problem that uses the previously estimated filter coefficients to insert the new, interpolated, traces between the original ones. We extended this two-step strategy to data interpolation by employing a device in which we used filter coefficients from a denser variable offset gather to interpolate the missing traces on a few independently constructed gathers. We applied the methodology on synthetic and real data sets, the latter acquired in the interior of the Antarctic continent. The variable-offset interpolated data opened the door to prestack processing, making feasible the production of a prestack time migrated section and a 2D velocity model for the entire profile. Notwithstanding, we have used a data set obtained in Antarctica; there is no reason the same methodology could not be used somewhere else.

We focus here on three horizons conspicuously embedded in the rich radar stratigraphy revealed on the fixed-offset radar data obtained Plateau Detroit, Antarctic Peninsula. Spatial filtering removed the more energetic reflection field and the surface wave arrivals at the earlier time, leaving only the diffracted field. This is particularly striking for the early time horizon where the direct wave arrivals had shrouded the diffractions before filtering. The density estimates and the photographic datasets from a centrally located well at depths compatible with the diffraction horizons suggested they share a common origin: a vertical transfer of mass associated with the formation of surficial hoar from a strong vertical temperature gradient in the snow cover, followed by a quick burial by fresh snow in a high accumulation environment. We have inverted the fundamental mode of the phase velocity dispersion of the surface waves to obtain a group velocity estimate and its depth range, used to improve the 1–D velocity model from a CMP gather by correcting its first velocity estimate. The same inversion solved an apparent ambiguity in our data by associating the surficial horizon with a specific density residual. We have also shown through modeling that the diffraction horizons seen in our data can be explained by the existence of large coarse–grained faceted crystals which became denser with depth than the surrounding firn. Keywords: GPR, radar, diffraction field, velocity model, waveguide, polar stratigraphy, depth hoar, Antarctica. RESUMO. Concentramo-nos aqui em três horizontes conspicuamente embutidos na rica estratigrafia revelada nos dados de radar de afastamento constante obtidos Platô Detroit, Península Antártica. Uma filtragem espacial removeu o campo de reflexão mais energético e as chegadas de onda de superfície das primeiras chegadas, deixando apenas o campo difratado. Isso é notável para o horizonte de primeiras chegadas, onde as ondas diretas encobriam as difrações antes da filtragem. As estimativas de densidade e os conjuntos de dados fotográficos do poço localizado no centro da aquisição mostra em profundidades cristais compatíveis com os horizontes de difração dos dados de GPR sugerindo uma origem comum: uma transferência vertical de massa associada à formação de hoar devido a um forte gradiente vertical de temperatura na cobertura de neve fresca em um ambiente de alta acumulação. Nós invertemos o modo fundamental da dispersão de velocidade de fase das ondas de superfície para obter uma estimativa de velocidade de grupo e sua faixa de profundidade usada para melhorar o modelo de velocidade 1-D a partir de um CMP, corrigindo sua primeira estimativa de velocidade. A mesma inversão resolveu uma ambiguidade aparente em nossos dados ao associar o horizonte superficial a uma densidade específica residual. Também mostramos através de modelagem que os horizontes de difração observados em nossos dados podem ser explicados pela existência de grandes cristais facetados de granulação grossa que se tornaram mais densos com a profundidade do que o firn circundante. Palavras-chave: GPR, radar, campo de difratado, modelo de velocidade, guia de onda, estratigrafia polar, depth hoar, Antarctica

Traditionally, first-arrival travel-time tomographic problems are treated as nonlinear inverse problems solved by deterministic optimization methods. These methods minimize the objective function by taking its derivatives. In some problems these derivatives are slow to compute and the method becomes inefficient. Another limitation inherent to deterministic methods is the strong relation between the starting model and local minima entrapment. To deal with these drawbacks, we apply a genetic algorithm with elitism (EGA) to cross-hole tomography of Ground Penetrating Radar (GPR) data. Regardless the starting model, this heuristic method ideally finds the region of the solution space containing global minima without calculating derivatives. We obtain good results in applying our method in two noise-corrupted synthetic cross-hole data sets. The first simulates horizontal strata and the second consists of a homogeneous background medium with single low-slowness anomaly. The methodology is also applied to field data where a cross-strata pattern is achieved. This pattern is in accordance with the geological feature of the region. Heuristic methods applied to high-dimensional geophysical problems yield good results but are computationally expensive. Coupling heuristic and deterministic methods is recommended to exploit the full power of each optimization technique.