A positive degree-day (PDD) model was tested on Antizana Glacier 15α (0.28 km<sup>2</sup>; 0°28' S, 78°09' W) to assess to what extent this approach is suitable for studying glacier mass balance in the inner tropics. Cumulative positive temperatures were compared with field measurements of melting amount and with surface energy balance computations. A significant link was revealed when a distinction was made between the snow and ice comprising the glacier surface. Significant correlations allowed degree-day factors to be retrieved for snow, and clean and dirty ice. The relationship between melt amount and temperature was mainly explained by the role of net shortwave radiation in both melting and in the variations in the temperature of the surface layer. However, this relationship disappeared from June to October (Period 1), because high wind speeds and low humidity cause highly negative turbulent latent heat fluxes. However, this had little impact on the computed total amount of melting at the annual time scale because temperatures are low and melting is generally limited during Period 1. At the daily time scale, melting starts when daily temperature means are still negative, because around noon incoming shortwave radiation is very high, and compensates for energy losses when the air is cold. The PDD model was applied to the 2000–2008 period using meteorological inputs measured on the glacier foreland. Results were compared to the glacier-wide mass balances measured in the field and were good, even though the melting factor should be adapted to the glacier surface state and may vary with time. Finally, the model was forced with precipitation and temperature data from the remote Izobamba station and NCEP-NCAR reanalysis data, also giving good results and showing that temperature variations are homogenous at the regional scale, meaning glacier mass balances can be modelled over large areas.