Estimates of daily surface meltwater production across the Antarctic Peninsula over 2007-2022. These data were generated using a random forest model combining enhanced resolution ASCAT C-Band radar backscatter, ASCAT-detected surface melt presence (https://zenodo.org/record/7995998), and ERA5 or MERRA-2 reanalysis sensible and latent heat fluxes, downward shortwave and longwave radiation, and 2-m temperature. Initial model training was undertaken using meteorological observations and surface melt estimates from surface energy balance modeling at several Larsen ice shelf automated weather stations (Jakobs et al., 2020), before scaling up to the Antarctic Peninsula region using ERA5 or MERRA-2. Note that the ASCAT-ERA5 melt estimates here are updated to v2 using a revised random forest melt prediction model. The ASCAT-MERRA-2 melt estimates here are the initial release (v1). This record will be updated upon publication of a forthcoming manuscript. Please reach out to Luke Trusel with questions!