The spread of ground source heat pumps has led to the development of several models to estimate the thermal interference between independent geothermal boreholes in densely populated areas. Some of these models take into account the changes in the ground surface temperature, thus allowing to account for the heat flux from the buildings, potentially high in densely populated areas. In this study, we investigated two models accounting for the ground surface temperature change; both models are based on the Stacked Finite Line Source (SFLS) model. The first model treats the ground surface as a set of rectangles, each with its own temperature; the second model treats the ground as an infinite surface with imposed temperature. The first model is more accurate, as it allows to consider different temperatures for different surfaces. However, the second model requires less computational time and is easier to implement. Therefore, we investigated whether the latter model can provide a good approximation of the former. Our results suggest that the second model can often provide a good approximation with significantly less computational time, especially for larger built areas; however, it could be unacceptably inaccurate for smaller neighbourhoods. We also investigated a strategy to improve its accuracy without compromising its computational speed; the strategy seems promising but requires more investigation.