Abstract. Strong seasonal variability of hygric and thermal soil conditions are a defining environmental feature in northern Australia. However, how such changes affect the soil-atmosphere exchange of nitrous oxide (N 2 O), nitric oxide (NO) and dinitrogen (N 2 ) is still not well explored. By incubating intact soil cores from four sites (three savanna, one pasture) under controlled soil temperatures (ST) and soil moisture (SM) we investigated the release of the trace gas fluxes of N 2 O, NO and carbon dioxide (CO 2 ). Furthermore, the release of N 2 due to denitrification was measured using the helium gas flow soil core technique. Under dry pre-incubation conditions NO and N 2 O emissions were very low (< 7.0±5.0 µg NO-N m −2 h −1 ; < 0.0±1.4 µg N 2 O-N m −2 h −1 ) or in the case of N 2 O, even a net soil uptake was observed. Substantial NO (max: 306.5 µg N m −2 h −1 ) and relatively small N 2 O pulse emissions (max: 5.8 ± 5.0 µg N m −2 h −1 ) were recorded following soil wetting, but these pulses were short lived, lasting only up to 3 days. The total atmospheric loss of nitrogen was generally dominated by N 2 emissions (82.4-99.3 % of total N lost), although NO emissions contributed almost 43.2 % to the total atmospheric nitrogen loss at 50 % SM and 30 • C ST incubation settings (the contribution of N 2 at these soil conditions was only 53.2 %). N 2 O emissions were systematically higher for 3 of 12 sample locations, which indicates substantial spatial variability at site level, but on average soils acted as weak N 2 O sources or even sinks. By using a conservative upscale approach we estimate total annual emissions from savanna soils to average 0.12 kg N ha −1 yr −1 (N 2 O), 0.68 kg N ha −1 yr −1 (NO) and 6.65 kg N ha −1 yr −1 (N 2 ). The analysis of longterm SM and ST records makes it clear that extreme soil saturation that can lead to high N 2 O and N 2 emissions only occurs a few days per year and thus has little impact on the annual total. The potential contribution of nitrogen released due to pulse events compared to the total annual emissions was found to be of importance for NO emissions (contribution to total: 5-22 %), but not for N 2 O emissions. Our results indicate that the total gaseous release of nitrogen from these soils is low and clearly dominated by loss in the form of inert nitrogen. Effects of seasonally varying soil temperature and moisture were detected, but were found to be low due to the small amounts of available nitrogen in the soils (total nitrogen < 0.1 %).

Abstract. Strong seasonal variability of hygric and thermal soil conditions are a defining environmental feature in Northern Australia. However, how such changes affect the soil–atmosphere exchange of nitrous oxide (N2O), nitric oxide (NO) and dinitrogen (N2) is still not well explored. By incubating intact soil cores from four sites (3 savanna, 1 pasture) under controlled soil temperatures (ST) and soil moisture (SM) we investigated the release of the trace gas fluxes of N2O, NO and carbon dioxide (CO2). Furthermore, the release of N2 due to denitrification was measured using the helium gas flow soil core technique. Under dry pre-incubation conditions NO and N2O emission were very low (<7.0 ± 5.0 μg NO-N m−2 h−1; <0.0 ± 1.4 μg N2O-N m−2 h−1) or in case of N2O, even a net soil uptake was observed. Substantial NO (max: 306.5 μg N m−2 h−1) and relatively small N2O pulse emissions (max: 5.8 ± 5.0 μg N m−2 h−1) were recorded following soil wetting, but these pulses were short-lived, lasting only up to 3 days. The total atmospheric loss of nitrogen was dominated by N2 emissions (82.4–99.3% of total N lost), although NO emissions contributed almost 43.2% at 50% SM and 30 °C ST. N2O emissions were systematically higher for 3 of 12 sample locations, which indicates substantial spatial variability at site level, but on average soils acted as weak N2O sources or even sinks. Emissions were controlled by SM and ST for N2O and CO2, ST and pH for NO, and SM and pH for N2.