The objective of the proposed research is to develop and validate a practical framework for transmitter localization and spectrum reconstruction that remains accurate under very sparse sensing and realistic geometric uncertainty. Three gaps that limit current spectrum cartography will be addressed: first, a lack of evidence on when higher fidelity propagation models materially improve results over simple path loss models in sparse regimes; second, the absence of calibrated uncertainty estimates that reflect errors in environmental geometry; and third, the prevailing assumption of dense, geographically distributed sensor data, which is often impractical or unavailable in operational deployments. Two focused contributions are proposed. First, benchmarks will be established across three model tiers, namely empirical log distance with shadowing, deterministic Terrain Integrated Rough Earth Model, and site specific ray tracing, on The Platform for Open Wireless Data-driven Experimental Research (POWDER) testbed using fixed and mobile endpoints with known ground truth. Metrics to be evaluated will include transmitter localization error and power estimation error, radio map root mean square error, calibration of uncertainty expressed as coverage of 90 and 95 percent intervals, and computational cost. Second, an ensemble geometry perturbation method will be developed that injects measured uncertainties in building height, position, and plausible material variability into the inference pipeline to produce confidence bounds for location and field estimates. The goal of this research is to obtain fundamental insights into the trade offs between sensing density, propagation model sophistication, uncertainty quantification, and spectrum intelligence performance that will inform the design of next generation autonomous spectrum management systems with robust uncertainty aware decision making capabilities for Radio Dynamic Zones and for 5Gand 6G network optimization.