Abstract:
Soil system functioning is essential for both humans and the environment. Soils are involved in all biogeochemical cycles, e. g., the carbon or the nitrogen cycle, thereby fulfilling their key function of cycling and retaining nutrients that are used by natural ecosystems as well as for agriculture. The application of fertilizer and pesticides during agricultural use puts soil functions under a constant stress. While soils can help in attenuating micropollutants or nitrate and thereby prevent them from leaching into the groundwater, soil functions can be impaired by the application of pesticides or fertilizer. Another stressor for soils is climate change, which enhances the frequency and intensity of droughts and rainfall events.
The goal of this thesis was to investigate the effects of a heavy rainfall event after a summer drought on soil redox potentials, inorganic nitrogen turnover, vertical transport of water, nitrate, and glyphosate and the soil microbial community in a fine-textured floodplain soil. It was hypothesized that (i) shrinkage cracks forming during the drought period can act as preferential flow paths thereby allowing dissolved and particle-bound compounds to leach into the subsoil during a heavy rainfall event, (ii) upon rewetting of the dry soil a pulse of inorganic nitrogen is observed (“Birch effect”), which, after a few days, transitions to the “usual” redox-driven nitrogen turnover processes, and (iii) the rewetting of the dry soil impacts the soil microbial community, especially its active fraction.
In preparation for the field study, the methodology for the quantification of nitrate and ammonium concentrations in soil samples was reviewed by testing the stability of samples during storage and extraction under different conditions (storage temperature, extraction time and extraction solution, oxygen availability). In summary, it is recommended, to freeze field soil samples immediately after sampling, to store them frozen and to extract them with 2 M KCl to avoid any impairment of ammonium and nitrate concentrations.
At a floodplain in the catchment of the Ammer river close to Tübingen, SW Germany, a field experiment was conducted to test the above listed hypotheses by simulating a heavy rainfall event on a dry soil. The floodplain soil was a fine-textured Gleysol, that exhibited pronounced shrinkage cracks after a summer drought period. On a harvested summer barley acre, the field experiment setup with nine equally-sized plots was installed. Three plots each were treated similar: (i) Plots received the simulated heavy rainfall event (deuterium-labeled water) and a light daily irrigation for 10 consecutive days. (ii) Plots were sprayed with glyphosate one day before the simulation of the heavy rainfall event as described in (i). (iii) Dry plots received no treatment and served as control plots. 50 cm soil cores were retrieved from all plots two hours after the simulated heavy rainfall event (i. e., day 0), on days 6/8 and 14. Analyses of the pore water content, the δ2H-signature of the pore water, glyphosate, ammonium and nitrate concentrations, the microbial community on DNA and RNA level, and extracellular enzyme activities were conducted. Additionally, in situ soil redox conditions were monitored in different soil depths.
The role of shrinkage cracks as preferential flow paths was evaluated by the δ2H-signature of the pore water, the isotope mass balance and the water mass balance, which consistently showed that already on day 0 irrigation water had reached depths down to 50 cm below ground level, i. e., the maximum depth investigated. Also glyphosate was found alongside with elevated isotopic signatures over entire core profiles. As glyphosate is strongly sorbing to mineral particles, this finding is surprising and indicates that a particle-bound transport of glyphosate through preferential flow paths must have occurred. Nitrate as very soluble compound is also prone to transport along preferential flow paths. Although a loss of nitrate in the topsoil was observed upon the heavy rainfall event, no nitrate accumulation was observed in the subsoil. As, however, other processes of nitrate loss could be excluded, it was assumed that the nitrate leached even further down than 50 cm below ground level and was therefore not recovered.
The assessment of nitrogen turnover processes occurring upon the soil rewetting by the simulated heavy rainfall event revealed: (i) Upon the initial rewetting of the soil during the simulated heavy rainfall event, a pulse of inorganic nitrogen in the form of ammonium appeared in the top ten centimeters of the soil (Birch effect). Except from that, ammonium behaved very conservatively, which can be explained by the high fraction of sorbed ammonium having a limited bioavailability compared to exchangeable ammonium. (ii) In situ redox potentials dropped from oxidizing to reducing within hours to days in response to the simulated heavy rainfall event and stayed low especially in the subsoil and controlled the inorganic nitrogen turnover processes in later stages of the field experiment (days 6 and 14). (iii) Nitrate contents responded to changes of in situ redox potentials, with a denitrification-related decrease at reducing conditions and a nitrification-driven increase at oxidizing conditions.
The effect of the simulated heavy rainfall event on the soil microbial community composition on both DNA and RNA (active fraction) level was negligible. However, lower diversity indices of the active fraction were measured on rewetted plots compared to dry plots, which indicated a certain selection for taxa better adapted to dry-wet-cycles.
The presented thesis demonstrates that the methodology of the described field experiment was suitable to study several interacting processes during the rewetting of a dry soil during a (simulated) heavy rainfall event. The insights promote the understanding of these interacting processes and the protection of soil and groundwater resources during drought periods and heavy rainfall events and lay a profound basis for further studies.