Ambient gas-phase and snow-phase measurements of formaldehyde (HCHO) were conducted at Summit, Greenland, during several summers, in order to understand the role of air-snow exchange on remote tropospheric HCHO and factors that determine snowpack HCHO. To investigate the impact of the known snowpack emission of HCHO, a gas-phase model was developed that includes known chemistry relevant to Summit and that is constrained by data from the 1999 and 2000 field campaigns. This gas-phase-only model does not account for the high ambient levels of HCHO observed at Summit for several previous measurement campaigns, predicting approximately 150 ppt from predominantly CH4 chemistry, which is ~25--50% of the observed concentrations for several years. Simulations were conducted that included a snowpack flux of HCHO based on HCHO flux measurements from 2000 and 1996. Using the fluxes obtained for 2000, the snowpack does not appear to be a substantial source of gas-phase HCHO in summer. The 1996 flux estimates predict much higher HCHO concentrations, but with a strong diel cycle that does not match the observations. Thus, we conclude that, although the flux of HCHO from the surface likely has a significant impact on atmospheric HCHO above the snowpack, the time--dependent fluxes need to be better understood and quantified. It is also necessary to identify the HCHO precursors so we can better understand the nature and importance of snowpack photochemistry.