GNGTS 2014 - Atti del 33° Convegno Nazionale

118 GNGTS 2014 S essione 3.1 Vorst, 1987). The space is discretized into layers, with dipping or curved separating interfaces, constituted by pixels (voxels in 3D), with homogenous velocity. Through a minimization of the time residuals, it is possible to invert the data both for velocities and for the interface position and shape (e.g. Vesnaver et al. , 2000; Rossi et al. , 2007). Reflected events provide information on the velocity of the layer above the reflected interface, whereas the refracted arrivals give information also on the velocity of the layer underlying the refracting interface, in correspondence of the ray-path part travelling along the interface. For both reflected and refracted arrivals, it is possible to exploit the time-residuals, converted to depth, to reconstruct the interface depth and shape (e.g., Rossi et al. , 2001, 2007). About line L100, for the moment, we identified reflected and refracted arrivals from three different interfaces. The results obtained are still preliminary, since they have to be confirmed and integrated with the data from the other two lines. They are, however, promising. As said, the model is constituted by four layers, separated by three interfaces. The approximate depth of the latter is about 15 m, 30 m and 55 m below the topographic surface respectively. Velocities lie within the 1.55 to 4.6 km/s interval, in agreement with the observations fromboreholemeasurements and 2D seismic lines in the sameAdventdalen valley, near CO2Clab (Oye et al. , 2013). There is evidence of strong velocity variations, both laterally as well as in depth, as it is expected across the pingo. Conclusions. From the very first results of the analysis of the data acquired during the IMPERVIA spring 2014 experiment, we can conclude that seismic method can be applied also for near surface, high resolution targets in presence of permafrost. The observations and the results of the field tests on the sensors and sources and of the successive analysis of the out- coming data can provide guidelines for successive surveys in similar areas. In particular, it is necessary to guarantee a good ground coupling, for both sources and receivers, by digging snow pits. It is also important to reduce the noise, due to the frequent strong wind, by burying the sensors, and covering with snow also the streamer connecting cables (e.g. Lecomte et al. , 2014). The choice of the source has to be done, so to have energy also at the long distances required for a correct velocity retrieval. Off-end shooting and cross-line recording can help to Fig. 3 – a) An example of shot gather from L300 (inline) and L200 (crossline) usig the Seisgun as source. b) The same using firecrackers as source.