Modern agriculture relies on a set of technologies which became implemented on large scale during the Green Revolution between the 1940s and the late 1960s; among these technologies the use of synthetic nitrogen fertilizers plays a key role in achieving high yields which are necessary to sustain the high (and growing) world population. However, the introduction of anthropic reactive nitrogen into the global nitrogen cycle also causes adverse environmental effects, such as the eutrophication of water bodies and the emission of nitrous oxide (N2O) into the atmosphere, where this gas contributes to global warming and to the depletion of stratospheric ozone.
Soils are a major source of N2O: in this environment the gas is produced mainly by an aerobic microbial process (nitrification) and by an anaerobic microbial process (denitrification), both of which are strongly affected by environmental drivers such as soil type, soil water content, soil temperature, soil nitrogen content, soil pH and agricultural management practices. These environmental drivers interact with each other in complex ways and this results in very high spatial and temporal variability in N2O emissions. This, in turn, makes the precise quantification of emissions (necessary in order to compare the effectiveness of mitigation strategies) and their prediction by means of mathematical models (necessary in order to implement the knowledge acquired on mitigation strategies under different sets of environmental conditions) very challenging.
In collaboration with partners within the NORA project, the YARA Research Centre Hanninghof is conducting experiments aimed at improving quantification of N2O emissions, at testing mitigation options and at improving mathematical simulation models.
In 2013 winter barley was cultivated in a field close to the city of Münster (north-western Germany) and N2O emissions were monitored under conditions of i) no nitrogen fertilization, ii) 200 kg of nitrogen per hectare in the form of calcium ammonium nitrate (CAN) and, iii) 200 kg of nitrogen per hectare in the form of urea. The results show that before harvest the emissions were low (37 g of nitrogen in the form of N2O per hectare = 37 g N2O-N ha-1) in the unfertilized control plots and significantly higher in the two fertilized plots (826 and 691 g N2O-N ha-1 for CAN and urea respectively). In contrast, after harvest the emissions from the unfertilized plots (751 g N2O-N ha-1) were much closer to those from the fertilized plots (1000 and 993 g N2O-N ha-1). These data show that, under conditions representative for a large part of north-western Germany agricultural soils, nitrogen fertilization induced emissions mainly before harvest, that a large fraction of the emissions took place after harvest when emissions were largely independent from direct nitrogen inputs, and that the uncertainty in the quantification of total emissions made the observed difference between the fertilizer types not statistically significant. The data also show that emissions were highly variable in time, and that this largely contributed to the uncertainty which makes the comparison among treatments so challenging.
The measured N2O emissions were compared with those calculated by the model DNDC95, one of the most widespread mathematical models capable of simulating nitrous oxide emissions: the comparison shows that the model was able to predict the time during which the main emission event of the year took place and that in general the magnitude of the maximum daily fluxes during the main emission events was accurately simulated. However, the model did not accurately simulate prolonged emissions of moderate intensity, which accounted for approximately half of the total emissions in the fertilized plots, and in general largely underestimated emissions.
These results highlight the importance of increasing the amount and quality of information on the effectiveness of N2O emission mitigation strategies and of improving the capability of models of accurately simulating measured emissions.