Microhabitat Effects on Nitrous Oxide Emissions, Production Pathways, and Reduction in Floodplain Soils

The potential of river floodplains to emit nitrous oxide (N2O), a powerful greenhouse gas and ozone-depleting compound, considerably reduces the climate regulation function of these dynamic transition zones between aquatic and terrestrial ecosystems. However, the assessment of N2O emissions from floodplain soils is challenging, due to the inherently high spatial heterogeneity and the characteristic occurrence of sporadic inundation phases. Such short-term flood events can temporarily alter the conditions for nitrogen (N) transformation processes taking place within distinct microhabitats, which can lead to the local formation of transient hot spots of enhanced N2O emissions. This situation emphasizes the urgent need to understand how characteristic factors of microhabitat formation in river floodplains control the balance between major microbial N2O source processes and N2O reduction to N2 that determine the magnitude and duration of N2O emissions. Therefore, the main objective of this thesis project was to systematically assess the relative importance of microhabitat effects related to soil aggregate size, organic matter accumulation, and plant-soil interactions on the microbial N2O production and consumption processes controlling the spatiotemporal emission patterns of N2O under changing pore water saturation. To achieve this objective, a mesocosm experiment under controlled climatic conditions, and a study within the framework of a field manipulation experiment were conducted.
In the mesocosm experiment, presented in chapters 2 and 3, two model soils with equivalent structure and texture, comprising macroaggregates (4000–250 μm) or microaggregates (< 250 μm) from a N-rich floodplain soil were used. These model soils were planted either with basket willow (Salix viminalis L.), mixed with leaf litter, or left unamended. The resulting six aggregate size / amendment factor combinations were exposed to a flood of 48 hours and then left to dry. Emission rates of N2O during the experiment were determined using the closed-chamber method. The relative contributions of different N transformation processes to the production of N2O and the degree of N2O reduction to N2 were assessed using a novel approach based on the isotopic and isotopomeric composition of the emitted N2O. In addition, the procaryotic and fungal soil microbiomes were characterized by sequencing of DNA and qPCR of functional genes related to potentially N2O producing and consuming processes. N2O production during the 48-hour flood phase originated almost entirely from heterotrophic bacterial denitrification and/or nitrifier-denitrification in all experimental treatments, yet most of the produced N2O was further reduced to N2 resulting in low N2O flux rates. In the drying phase, a period of enhanced N2O emissions occurred in all treatments, however with the unamended and litter added model soils with macroaggregates emitting significantly more N2O than in all other treatments. During this period, most of the N2O production continued to derive from bacterial denitrification in anoxic micro-sites. However, the aeration of the inter-aggregate pore space led to additional contributions by oxidative N2O production, the magnitude of which depended on treatment and time point within the drying phase. Also here, aggregate size emerged as a key parameter. Unamended macroaggregates seemed to prolong anoxia within microsites when compared to microaggregates. Litter addition further enhanced soil anoxia but also altered soil structure and nutrient availability. This increased soil heterogeneity modulated the temporal pattern of the N2O emission, leading to short-term peaks of high N2O fluxes at the beginning of the period of enhanced N2O emissions. These maximum N2O emissions were the result of rapid changes in N2O source partitioning of nitrifying and denitrifying processes in combination with a temporary partial disruption of N2O reduction. By contrast, the presence of S. viminalis prevented the occurrence of strong N2O emissions from both model soils, attenuating any effect of flooding and aggregate size on N2O production pathways and the degree of N2O reduction. Root respiration and the decomposition of root exudates likely promoted the formation of anoxic microsites that support complete denitrification, resulting in the low emission rates observed in the planted model soils. Irrespective of treatment and throughout the experiment, nitrogen-cycling gene abundances revealed a higher potential for bacterial denitrification and for N2O reduction than for ammonia oxidation, thus supporting the implications of the isotopic data on the dominance of denitrification in N2O production and on the generally high degree of N2O reduction. DNA sequencing data and functional gene abundances further revealed that large and small soil aggregates represent distinct microhabitats with a different potential for both denitrifying and nitrifying processes, thus suggesting that in addition to structure-related physical effects, differences in the microbial community composition contribute to aggregate size effects on N2O emission rates and N2O production pathways. Litter accumulations strongly altered the soil microbial community composition of the aggregate size fractions, whereas the presence of willow had little respective effects. Both soil amendments affected the abundance of ammonia-oxidizing bacteria and archaea, but not the one of denitrifying microorganisms.
The field study, presented in chapter 4, took place in the hydrologically most dynamic floodplain zone of a re-naturalized section of the Thur River in NE Switzerland. In a randomized complete block design, experimental plots where the dominant vegetation, the pioneer plant canary ryegrass (Phalaris arundinacea L.), was constantly removed were compared to unmanipulated plots. During a three-week drying-phase after a major flood, the dynamics of efflux, source partitioning and reduction of N2O were assessed using the same methods as in the mesocosm experiment. In addition, temporal changes in the activity of specific groups of N transforming soil microorganisms were analyzed using extracted RNA. It became evident, that young, sandy sediments under a dense plant cover experienced longer periods of elevated N2O emissions, whereas emissions from bare sediments gradually decreased after initial peak rates. Nitrification and/or fungal denitrification contributed consistently about 20-30 % to gross N2O production in plots covered by P. arundinacea, whereas this process group contributed only to the beginning of the post-flood phase to N2O emissions from bare plots. N2O reduction was temporarily interrupted at the beginning of the post-flood phase in bare plots, whereas N2O reduction in the Phalaris plots was stable during the entire drying phase. The detection of denitrifying and nitrite oxidizing microorganisms as the most active N transforming microorganisms in this part of the river floodplain further supported the results from source partitioning and reflected the adaptation of the microbial community to fluctuating redox conditions.
Overall, the results of this thesis (i) demonstrate the importance of soil aggregation, litter accumulation and plant-soil interactions in floodplain soils in governing the production, consumption, and emission of N2O during flood-induced hot moments, (ii) present evidence for the formation of related specific microhabitats and indications of explanatory physical effects, (iii) highlighted the role of microbial N2O reduction as a major controlling factor of N2O emissions (iv) and confirm the dominance of denitrifying processes as source processes. Our findings thus should help to predict the location of temporary hotspots of N2O emissions, and to improve the estimations of local N2O budgets of river floodplains in a world of global climate change.