Although Chemical Transport Models (CTMs) such as the Community Multiscale Air Quality Model (CMAQ) have been used in linking observations of trace gases to emissions and developing vertical column distributions, there remain consistent biases between CTM simulations and satellite retrievals. Simulated tropospheric NO vertical column densities (VCDs) are generally higher over areas with large NO sources when compared with retrievals, while an opposite bias is found over low NO regions. Artificial (i.e., numerical) dilution in the model, where emissions are mathematically dispersed uniformly within the originating CTM grid, can impact modeled NO:NO ratios, while lower CTM VCD levels often observed over rural areas can be attributed to missing emission sources of NO or flawed horizontal/vertical transport. Potential causes of both low and high biases are assessed in this study using CMAQ and Tropospheric Monitoring Instrument (TROPOMI) NO retrievals. It was found that more detailed modeling of NO plumes to assess the NO:NO ratio in two power plant plumes can mitigate the effect of artificial computational dilution, reducing the bias and overall differences in the observed vs modeled plumes (errors reduced by 30%). Adjustments of upper tropospheric NO led to overall improvements, with a reduction in CMAQ bias (-43% to -29%) and improved spatial correlation (0.81 to 0.86). This study highlights the importance of having accurate modeled NO:NO ratios when comparing models to retrievals and the impact of unintentional numerical dilution.