This study aims to assess the accuracy of a Bayesian inference-based Source Term Estimation (STE) algorithm in determining the location and release rate of a passive pollutant from an unknown stationary source in an urban environment under varying conditions. Specifically, this work extends the study by Gkirmpas et al. (Atmosphere 15:871, 2024 [1]) by further investigating how increasing the number of sensors in the measurement network affects the accuracy of the existing STE methodology. The Metropolis–Hastings Markov Chain Monte Carlo (MCMC) algorithm is used to estimate unknown source parameters as Probability Density Functions (PDFs), utilizing concentrations derived from a Computational Fluid Dynamics (CFD) model and observational data from the measurement network. Several release scenarios involving different source locations, wind directions, and synthetic observation datasets are explored within the domain of Augsburg city center. The findings suggest that the algorithm consistently delivers accurate results across all scenarios. Furthermore, the results show that in cases where only a limited number of sensors detect pollutant concentrations above a threshold value, increasing the number of sensors directly influenced by the pollutant plume can significantly improve the algorithm's accuracy. Conversely, adding sensors to an already dense network influenced by the plume may result in better, similar, or slightly worse estimations.