Accurate localization methods are a fundamental component of integrated sensing and communication (ISAC) systems, where sensing and positioning functions must coexist with communication tasks under tight resource constraints. In this work, we provide a comparative study of several classical and low-complexity strategies for two-dimensional (2D) localization with elevated uniform rectangular array (URA) by exploiting known communication pilot signals, a foundational scenario in ISAC. The evaluated techniques include standard MUSIC NF, 2D MUSIC, Adaptive Grid Search (Zoom-in), root-MUSIC, 2D ESPRIT, PEACH-MUSIC, DFT-NF beamforming, OMP NF and a bicubic interpolation refinement of the MUSIC pseudospectrum. The considered system consists of a single ground-level user and an elevated URA under free-space line of sight (LOS) propagation. The methods are assessed in terms of localization accuracy root mean square error (RMSE) versus signal-to-noise ratio (SNR) dB, localization accuracy (RMSE) versus number of snapshots and computational complexity (Mean execution time), highlighting the trade-offs between RMSE × SNR and execution time. Results demonstrate that while classical MUSIC achieves high precision at the expense of prohibitive search complexity, tested search-free methods such as Root-multiple signal classification (MUSIC) and 2D-ESPRIT exhibit performance degradation due to near-field (NF) mismatch but provide competitive accuracy with substantially reduced runtime in compatible conditions. Bicubic interpolation improves the coarse MUSIC grid estimates with negligible overhead, and PEACH-MUSIC demonstrates the lowest runtime among the evaluated algorithms in favorable SNR conditions, while its performance degrades at lower SNR due to reduced spatial diversity. These findings offer a quantitative comparison to guide the selection of localization algorithms where performance and complexity must be carefully balanced, a common requirement in ISAC-oriented architectures.