With the aim to investigate the production and destruction mechanisms of formaldehyde (HCHO) on a seasonal scale in the remote troposphere, HCHO was monitored over a 13-months period at Dumont d'Urville station, located near the East Coast of the Antarctic continent. Monthly means of 50 ppt in winter and up to 200 ppt in summer were measured. In spring, summer and autumn, a daily variation was detected with amplitude of ≈40 ppt and a maximum during the afternoon. The analytical instrument (Aerolaser, AL4021) deployed for the measurements, was found to be extremely sensitive to ambient temperature variations. The origin of the temperature dependence was investigated, and the deployed analytical method was adapted to eliminate this artefact. The impact of local contamination sources at the measurement site was explored in detail. Anthropogenic combustions were found to have a limited impact (maximum 150 pptv on hourly means) versus a local emission source from ornithogenic soils, induced by the omni-presence of Adelie penguins on the site in summer (maximum 200 pptv on hourly means). From november to march, these sources can introduce an overestimation by a factor 5 of the daily HCHO amplitude, while the influence on the monthly averaged budget can account up to 100 ppt. The raw dataset was thoroughly filtered using local meteorological criteria to eliminate any potential contamination. HCHO measurements were compared to simulations made with a numerical box model developed on the base of the current understanding of the photochemical sources and sinks of HCHO. During summer, methane oxidation was found to be the major source of HCHO production, likely due to enhanced levels of oxidants such as hydroxyl radical (OH) or nitrogen oxide (NO) occurring at DDU since this coastal site is under strong continental influence. Snow emissions and HCHO formation via methyl-hydroperoxyde oxidation complete the summer budget (respectively 10 to 20% of the CH4 oxidation). In contrary to what was observed for the Weddell Sea area, halogens chemistry is probably unlikely to contribute significantly at Dumont d'Urville (<10% of the CH4 oxidation). Finally, the initially attended summer source originating from non-methane hydrocarbons oxidation was found to be inefficient and to contribute only with less than 5 % of the CH4 oxidation. During winter the snowpack source was found to be preponderant (70% of CH4 oxidation) versus the gas phase chemistry sources of HCHO. Model investigations, performed on a seasonal and daily scale have shown that gas phase processes together with deposition and snow emission have to be considered to achieve a correct reconstruction of the observations.
