To assess the influence of various heat transfer processes on the thermal structure of near-surface ice in Greenland's ablation zone, we compare in situ measurements with thermal modeling experiments. A total of seven temperature strings were installed at three different field sites, each with between 17 and 32 sensors and extending up to 21 m below the ice surface. In one string, temperatures were measured every 30 min, and the record is continuous for more than 3 years. We use these measured ice temperatures to constrain our modeling experiments, focusing on four isolated processes and assessing the relative importance of each for the near-surface ice temperature: (1) the moving boundary of an ablating surface, (2) thermal insulation by snow, (3) radiative energy input, and (4) subsurface ice temperature gradients below the seasonally active near-surface layer. In addition to these four processes, transient heating events were observed in two of the temperature strings. Despite no observations of meltwater pathways to the subsurface, these heating events are likely the refreezing of liquid water below 5-10 m of cold ice. Together with subsurface refreezing, the five heat transfer mechanisms presented here account for measured differences of up to 3 °C between the mean annual air temperature and the ice temperature at the depth where annual temperature variability is dissipated. Thus, in Greenland's ablation zone, the mean annual air temperature is not a reliable predictor of the near-surface ice temperature, as is commonly assumed.