*3.2. Numerical Models of the Floor*

Figure 14 shows the GPR B-scans acquired in the numerical models of the considered floor. Taking into account the complexity of the analyzed medium, the electromagnetic wave velocity was set as constant and equal to 10 cm/ns. The strategy applied did not take into account the heterogeneity factor. Therefore, the results were only approximations of the real subsurface geometry. However, this fact did not disable the proper interpretation of the results. In the model #1.1 (Figure 14a), the positions of the stone tiles were clearly identified based on the line patterns denoting the reflections from the vertical joints between tiles. The half hyperbolas at the ends of the tombstone were visible in the center of the scan. The lower faces of the tiles and the tombstone were located at the depths of 0.04 and 0.15 m, respectively. Some additional reflections with a lower intensity were observed below, indicating the further wave reflections from the analyzed elements (e.g., at the depth of 0.08 m below the tiles and 0.3 m below the tombstone). The difference between the tiles laid with and without the air gap was small, i.e., additional reflections below the tiles located above the air gap were insignificantly stronger. This observation might lead to the hypothesis that the detection of the small air gaps based on the GPR scanning could be problematic. For the model #1.2 (Figure 14b), the conclusions from the identification of the tombstone and the tiles were the same as for model #1.1. However, additional hyperbola patterns appeared, being the reflections from all singular bricks of the walls located below the edges of the tombstone. It is worth noting that the intensity of the hyperbolas decreased with the depth, which is a common relation observed for electromagnetic waves. Similarly, another brick wall was visible at the distance of 2.55 m. It is interesting to note that the walls seemed to be located at different depths. This apparent observation was the result of the assumption that the electromagnetic wave velocity was constant. The wave needed more time to travel though the tombstone (to reach the edge walls) compared with the stone tiles and the sand (above the additional

wall). In the model #1.3 (Figure 14c), the additional hyperbolas were observed below the tombstone, proving the presence of the rubble. The great amount of the inclusions made the interpretation of the scan more complicated. It is also worth noting that the deeper hyperbolas had a much lower intensity, thus the identification of inclusions located deeper was straitened. The interpretation of the reflections below the tombstone was far more difficult when considering the simultaneous presence of the rubble and the brick walls (model #1.4, Figure 14d). The superimposition of the hyperbolas denoting the rubble and the singular bricks made them undistinguishable. Therefore, it was not possible to state whether a certain hyperbola denoted the brick or any different kind of inclusion. To sum up, the results obtained from the numerical calculations gave some information about the possibilities and limitations of GPR scanning. The location of the tombstone and the tiles could be determined. The rubble below the tombstone was detectable as well as the brick walls; however, when these elements appeared together, the interpretation was complicated. What is essential is that the air gaps were difficult for the identification.
