Abstract. We use two snow models, the IMAU Firn Densification Model (IMAU-FDM) and SNOWPACK, to model firn characteristics in the Antarctic Peninsula (AP). We force these models with mass and energy fluxes from the Regional Atmospheric Climate MOdel (RACMO2.3p2) to construct a 1979–2016 climatology of AP firn density, temperature and liquid water content. A comparison with 75 snow temperature observations at 10 m depth and with density from 11 firn cores, suggests that both snow models perform adequately. In this study, we focus on the detection of so-called perennial firn aquifers (PFAs), that are formed when surface meltwater percolates into the firnpack in summer, is then buried by snowfall, and does not refreeze during the following winter. In 941 model grid points, covering ~ 28,000 km2, PFAs existed for at least one year in the simulated period, most notably in the western AP. At these locations, surface meltwater production exceeds 150 to 300 mm w.e. yr−1, with accumulation at least an order of magnitude larger. Most pronounced and widespread are PFAs modelled on and around Wilkins ice shelf. Here, both meltwater production and accumulation rates are sufficiently high to cause PFA formation in most years in the 1979–2016 period, covering a large part of the ice shelf. Other notable PFA locations are Wordie ice shelf, an ice shelf that has almost completely disappeared in recent decades, and the relatively warm northwestern mountain ranges of Palmer Land, where accumulations rates can be extremely large and PFAs are formed frequently. We find that not only the magnitude of melt and accumulation is important, but also the timing. If large accumulation events occur in the months following an above average summer melt event, this favours PFA formation in that year. Finally, we find that most PFAs are predicted near the grounding lines of the (former) Prince Gustav, Wilkins and Wordie ice shelves. This highlights the need to further investigate how PFAs may impact ice shelf disintegration events, in a similar way as supraglacial lakes do.
The flexural dynamics of melting ice shelves
