Radio environment maps (REMs) have been the subject of increasing research attention in the recent years due to their aptitude for characterization of the radio frequency (RF) spectrum. Their reconstruction in three-dimensional (3D) space has become an important consideration for resource allocation in modern communication networks. This work evaluates the performance of two prominent volumetric interpolation methods - inverse distance weighting (IDW) and Kriging, by calibrating them through real-world RF data collected in 3D space within an anechoic chamber. Through Monte Carlo (MC) simulations, the methods are calibrated using this data and validated two other datasets obtained in the respective indoor and outdoor real-world scenarios, for different amounts of training samples. The results are compared against the previous calibration of these two methods, and the improved gains in terms of accuracy are illustrated and explained. Normalized RMSE (NRMSE) of as little as to -7.1 dB is achieved.