Source determination of N 2 O has often been performed using stable isotope incubation experiments. In situ experiments with isotopic tracers are an important next step. However, the challenge is to distribute the tracers in the field as homogeneously as possible. To examine this, a bromide solution was applied as a stand-in tracer using either a watering can, a sprayer, or syringes to a relatively dry (25% gravimetric moisture content) or wet (30%) silt loam. After 1 h, samples were taken from three soil depths (0-10 cm), and analyzed for their water content and bromide concentration. The application with syringes was unsuccessful due to blocked cannulas. Therefore, further laboratory experiments were conducted with side-port cannulas. Despite a larger calculated gravimetric soil moisture difference with watering can application, more Brtracer was recovered in the sprayer treatment, probably due to faster transport of Brthrough macropore flow in the wetter conditions caused by the watering can treatment. The losses of Br-(33% for the watering can, 28% for the sprayer treatment) are equivalent to potential losses of isotopic tracer solutions. For application of 60 at% 15 NH 4 + , this resulted in theoretical enrichments of 44-53 at% in the upper 2.5 cm and 7-48 at% in 5-10 cm. As there was hardly any NO 3 in the soil, extrapolations for 15 NO 3 calculated enrichments were 57-59 at% in the upper 2.5 cm and 26-57 at% in 5-10 cm. Overall, no method, including the side-port cannulas, was able to achieve a homogeneous distribution of the tracer. Future search for optimal tracer application should therefore investigate methods that utilize capillary forces and avoid overhead pressure. We recommend working on rather dry soil when applying tracers, as tracer recovery was larger here. Furthermore, larger amounts of tracer lead to more uniform distributions. Further studies should also investigate the importance of plant surfaces.

Isotopic tracer methods using 15N or other isotopes provide insights into the sources and underlying N transformations leading to N2O emissions from agricultural soils. However, homogeneous labelling of naturally structured soil in the field is challenging since macropore flow must be avoided while the application must be performed in a limited timeframe. Therefore, we tested the infiltration pattern of several application methods and consequently developed drip application (DA) for larger scales using individual dropper bottles. We performed a proof of concept test, followed by an evaluation experiment in-situ with a manual sprinkler as control at an undisturbed grassland site using 15NH415NO3 (80 kg N ha-1, 10 at.%, 15 mm precipitation equiv.). 15N-NH4+ and 15N-NO3- recovery rates and corresponding correlation coefficients were calculated to examine horizontal and vertical homogeneity. The proof of concept test showed the negative effect of very dry topsoil on homogeneous infiltration. DA achieved significantly larger 15N recovery rates than sprinkler application and led to a more homogeneous horizontal label distribution. For DA, coefficients of variation of 15N recovery rates were smaller than with sprinklers for most depths, yet for both methods 15N recovery rates especially of 15N-NH4+ decreased vertically. Besides optimised label distribution, the DA method offers high flexibility in application patterns while offering reproducibility, feasibility and a reasonable application speed also at undisturbed sites at the plot scale. Moreover, DA causes no change in soil structure or soil diffusivity. Thus, the drip application method was found suitable for tracer application to field sites.