High-Sensitivity GNSS receivers are capable to acquire and track signals at low signal strengths, which enables receivers to capture signals reflected from buildings and other objects. These multipath and echo-only effects, which lead to GNSS receivers' inaccuracy in locations such as urban canyons are the enablers of the 2.5D mapping framework proposed in this paper. Several methods are employed to analyze GNSS signal propagation and estimate distances and heights of walls, from which satellite signals, that were captured as weak, had been reflected. Our experiments validate the approach and demonstrate that analysis of GNSS signal reflection can be the basis for rough 2.5D mapping. In turn, the technique can be used either as a standalone mapping solution, or in the context of Simultaneous Localization and Mapping (SLAM).