Since a few decades, the balance of the nitrogen (N) cycle has been deeply disturbed by human activies. The global impact of these activities on the N cycle can be described as a doubling of the transfer from the vast and unreactive atmospheric pool to biologically available forms (N fixation). The main sources responsible for the increase of reactive N emissions are the use of artificial fertilisers (NH3) as well as the combustion processes (NOx). Reactive N is then transformed, transported by both the atmosphere and the hydrosphere and, finally, deposited on both the terrestrial and the aquatic ecosystems with potentially strong impacts. Amongst the terrestrial ecosystems, the temperate forests are particularly sensitive to reactive N increases for various reasons: they are located near to strongly anthropised areas and are thus subjected to strong depositions; they are naturally N-limited and their biochemical cycle can indeed be strongly influenced by additional N, which leads to eutrophication and potential impacts such as soil acidification, nitrate leaching and losses of biodiversity within microorganisms, plants and fauna communities. Previous studies carried out in temperate forest ecosystems have shown that the soil, namely the soil organic matter, acts as a main deposited N sink for the ecosystem. However, biogeochemical reactions responsible for N retention in the soil are still not fully understood. Moreover, the majority of the concerned studies were conducted in acidic or hydromorphous soils and very little is known today about the fate of N deposition under other soil physico-chemical conditions. Consequently, the present research aims at filling some of the gaps described above and deals with two main topics concerning the retention of atmospherically deposited N in soils that are (1) the characterisation of the retention of atmospherically deposited N in the soil, in terms of both duration and quantity, more specifically, in a well drained calcareous soil and (2) the mechanisms and processes responsible for such a retention. From a practical point of view, the retention of N deposition in the soil was characterised by means of a 15N field labelling experiment simulating N atmospheric deposition and conducted at the Grandvillard research site, in the riparian zone of La Sarine River (Swiss Prealps). The stand is a beech forest mixed with planted spruces. The soil consists of a well drained calcareous fluvisol with a fast organic matter turnover. We tracked the N tracers (15NH4+ or 15NO3-, corresponding to the main forms of deposition) from short-term (hours, days or weeks) to longer-term intervals (one year), by measuring the partitioning of 15N into different biochemical soil fractions (extractable N, microbial N, roots N and N immobilised in soil). Aiming at studying the mechanisms and processes responsible for N retention, we performed two laboratory experiments. The first one was an acid hydrolysis, in order to determine the proportion of N deposition retained in the hydrolysable fraction (i.e. the more labile N forms, including the most bioavailable ones) as well as in the non-hydrolysable fraction covering the more recalcitrant N compounds. Within the framework of this experiment, we compared the results obtained in Grandvillard and in an aditional constrasting site (Alptal). The second experiment was a short-term (one hour to one week) laboratory incubation of sterilised and not sterilised soils, which purpose was to highlight the relative importance of biotic and abiotic processes in N immobilisation. Soils were labelled with either 15NH4+ or 15NO3- and subjected to 15N CPMAS NMR spectroscopy analysis, which is a powerful technique to identify the N organic molecules in soils. Within the framework of the field labelling experiment, more than half of the tracers (either 15NO3- or 15NH4+) was recovered in the soil between one hour and one year after labelling. Therefore, the general assumption that non-fertilised temperate forest soils are the main sink for deposited N can be confirmed. Since previous studies showing the importance of the soil for the retention of deposited N were all conducted on either acidic or hydromorphous soils, our results allow us to extend the validity of their general conclusions to a wider range of unfertilised temperate forest soils. We further showed that the main forms of N deposition (i.e. NO3-and NH4+) were retained in the soil within the same range of magnitude, in spite of their different biochemical pathways. N retention occurred mainly in the soil organic layers and, consequently, our results confirme the importance of the organic fraction in the deposited N retention. Since deposited N was retained in the soil in the very short term, we confirmed the presence of very fast processes and demonstrated that the very short term dynamics determined the fate of deposited N in the longer term: the main processes were i) a loss of deposited N as extractable through lixiviation (and lateral fluxes) and ii) a retention as N immobilised in soil. Furthermore, we demonstrated that N immobilisation brought deposited N both into the hydrolysable and into the recalcitrant fractions of soil N within short-time range (i.e. one week) . We also demonstrated that the hydrolysability (hydrolysable N/ total N) was constant over the year. Within the framework of the laboratory incubation of soils subjected to the 15N CPMAS NMR spectroscopy analysis, we showed that all NMR spectra were dominated by a single signal corresponding to the amide-peptide structure, whatever the soil layer concerned (organic or organo-mineral), the form added (NO3- or NH4+) and whether the soil was sterilised or not. Such dominance of proteinaceous compounds was in agreement with the results obtained in humic substances and in various soils all over the world. Further to the extractable 15N dynamics during the incubation, we proposed that biotic processes were dominant in the short-term N immobilisation. However, an abiotic fixation of less importance was not excluded for 15NO3-. Considering our results, we propose two mechanisms responsible for the retention of deposited N in soil in the long term. The first mechanism is a rapid and stable immobilisation in the recalcitrant pool of soil organic matter. Our results showed that biotic and abiotic processes could be involved: the afore-mentioned process of abiotic nitrate fixation could lead to the N immobilisation in heterocycles whereas N incorporated in amides and then stabilised in soil organic matter seemed to be the major pathway. The second mechanism of retention we are proposing is the biological recycling within the soil or the soil-plant system: on a short scale, microbial recycling seemed to be very rapid because deposited N was mainly present under a biosynthesised form (amides), in spite of the low amounts of tracer found in the microbes. Within the soil profile, roots recycling was probably efficient: fine roots, which are often the major contributors of soil organic matter were an important sink for N deposition all over the year. The combination of various scales and techniques allowed the connection between the dynamics of N retention in soil organic layers observed and the mechanisms responsible for it (i.e. rapid immobilisation in a recalcitrant pool and biological recycling through micro-organisms and plants). The results of this multi-scale approach suggest further developments and impacts for the modelling of the long-term fate of deposited N.