Determination of ammonia emissions in multi-plot field trials to assess optimized application techniques for liquid manure in growing winter wheatDissertation
Datum der mündl. Prüfung:2023-09-01
Betreuer:Prof. Dr. Klaus Dittert
Gutachter:Prof. Dr. Hans-Werner Olfs
Gutachter:Prof. Dr. Heinz Flessa
EnglischAmmonia emissions caused by liquid manure application affect human life expectancy and threaten natural ecosystems. However, other associated concerns like greenhouse gas emissions and nitrate leaching are equally relevant. Thus, German legislation severely restricted autumn application of liquid manure, since most crops have low nutrient demand at this development stage, so that much of the applied nitrogen would be lost to the environment. Therefore, liquid manures have to be applied in spring into growing crops. However, immediate incorporation into the soil to minimize ammonia emissions is not possible under these circumstances. Moreover, biogas digestate has become an increasingly popular organic fertilizer over the last three decades, since anaerobic fermentation is a climate friendly energy source. However, it might be associated with increased ammonia emissions due to its comparatively high pH and ammonium content. Therefore, liquid manure should be applied using optimized techniques for growing crops to mitigate ammonia emissions. Those techniques are based either on reducing the contact of fertilizer and atmosphere or on acidifying liquid manure. To evaluate optimized techniques, ammonia emissions have to be quantified in multi-plot field trials. Unfortunately, standard micrometeorological methods require large field areas and expensive equipment, making them difficult or even impossible to apply. Thus, other approaches adjusted to those specific requirements are used in multi-plot field trials. Calibrated passive sampling uses acid traps placed in the center of each plot to absorb ammonia, which enables a relative comparison of emissions. Subsequently, acid trap samplings are scaled by simultaneous measurements with the dynamic tube method, which uses a mobile chamber system to quantify ammonia emissions. The first objective of this study was therefore to evaluate calibrated passive sampling in multi-plot field trials with liquid manure application. However, ammonia drift between plots as well as chamber system contamination might be a particular challenge in such an experimental set-up. Therefore, the first step was to analyze the potential influence of the ammonia background on acid trap samplings and dynamic tube measurements. In a second step, the best practice to scale relative differences between plots obtained from acid trap samplings was assessed. In order to reduce costs and to minimize chamber system contamination, dynamic tube measurements are only performed on a few selected plots. Thus, characteristics of a well-suited treatment to perform simultaneous measurements with both methods were evaluated. However, the transfer coefficient (cumulated qualitative emissions divided by cumulated acid trap samplings) required to scale emissions might depend on the level of data aggregation. It can be calculated based on individual plots, treatment means or all plots of a field experiment. Therefore, it was evaluated which amount of data aggregation is sufficient. The second objective of this study was to evaluate optimized techniques to apply liquid manure in growing winter wheat in a series of field experiments in Germany. Calibrated passive sampling was used to assess ammonia emissions and yield and nitrogen uptake were measured as well to allow an agronomical evaluation of those techniques. Nitrogen fertilizer were applied at a total rate of 170 kg N ha−1 split into two equal dressings. Each experiment consisted of several techniques to apply cattle slurry and biogas digestate: i) trailing hose application using untreated and ii) acidified liquid manure, as well as iii) a combination of open slot injection for the first dressing and trailing shoe application for the second dressing. Furthermore, ammonia emissions, yield and nitrogen uptake of organically fertilized treatments were put into perspective by also implementing a treatment with mineral fertilization (broadcast calcium ammonium nitrate) and an unfertilized control. Furthermore, the unfertilized control was crucial to assess the influence of ammonia drift between plots. Acid trap samplings differed significantly between control plots, indicating that cumulated samplings of each individual plot depend not only on the ammonia emissions of the respective plot, but also on its specific background. Hence, many replications are necessary to obtain valid treatment means and those mean values show high standard deviations. However, there is no evidence, that passive sampler results are generally biased. Therefore, they are an easy way to obtain relative comparisons between treatment means. For the dynamic tube method, ammonia drift between plots had only a minor impact. However, we showed that chamber system contamination has a profound effect on calculating cumulated ammonia emissions in multi-plot field trials. The on field cleaning procedure using paper towels was not sufficient to reduce contamination. The relative influence of background ammonia was higher in treatments with low emissions for both methods. Therefore, scaling of acid trap samplings by simultaneous dynamic tube measurements should be performed in a treatment with high ammonia emissions. Regarding the amount of data aggregation required to scale emissions, this thesis showed that calculating a transfer coefficient based on individual plots is not sufficient, due to the influence of the fluctuating ammonia background. Therefore, acid trap samplings were scaled based on mean values in a treatment with high ammonia emissions. In this series of winter wheat field trials, the highest ammonia emissions (on average 24 kg N ha−1) occurred following trailing hose application. Applying biogas digestate lead to approximately 60 % higher emissions than cattle slurry application. Overall, acidification reduced emissions by 64 % for both liquid manure types. On average, the combination of slot injection and trailing shoe application resulted in 23% lower ammonia emissions compared to trailing hose application. However, decreasing ammonia emissions did not increase yield and nitrogen uptake. All treatments with liquid manure application led to similar crop yield (approximately 7 t ha−1 grain dry matter yield) and aboveground biomass nitrogen uptake (approximately 150 kg ha−1). Yield (8 t ha−1) and nitrogen uptake (approximately 190 kg ha−1) were significantly increased for the minerally fertilized treatment, while for the control, yield (approximately 4.5 t ha−1) and nitrogen uptake (approximately 90 kg ha−1) were significantly reduced. In summary, our results show that the mitigation of ammonia emissions originating from liquid manure application to growing crops is possible by using optimized application techniques. For this series of field trials, acidification was the technique with the highest ammonia mitigation potential. Future studies using calibrated passive sampling should address the importance of ammonia drift and chamber system contamination. Therefore, the use of separate dynamic tube chamber systems for each plot is recommended. Furthermore, increasing the plot size might reduce ammonia drift.
Keywords: ammonia emissions; organic fertilization; slurry acidification; trailing hose; trailing shoe; slot injection; calibrated passive sampling; dynamic tube method; acid traps; passive samplers; triticum aestivum; nitrogen use efficiency; biogas digestate; cattle slurry