Biosphere-Atmosphere Exchange of Peroxyacetyl Nitrate: Development of a Flux Measurement System and its Application on a Grassland Ecosystem

Peroxyacetyl nitrate (PAN) is an organic nitrogen species playing an important role in atmospheric chemistry. Being a source of other nitrogen oxides, it promotes for example the formation of aerosol particles and tropospheric ozone (O3). It has an impact on air quality, human health, as well as on aquatic and terrestrial ecosystems. Knowledge of the formation and removal of atmospheric PAN is thus important and also essential for the improvement of the chemistry of climate models. While PAN is known to be taken up by vegetation, underlying mechanisms are still not well understood and the role of PAN deposition as a removal process and source of nitrogen to ecosystems is not clear. This thesis presents a measurement system designed for the determination of biosphere-atmosphere exchange fluxes of PAN. The system was applied on a natural grassland ecosystem with the focus on the performance of the flux measurement system, but also with the aim to contribute to a better understanding of the biosphere-atmosphere exchange of PAN. The system was designed for the application of two flux measurement techniques, hyperbolic relaxed eddy accumulation (HREA) and the modified Bowen ratio (MBR) method, employing a gaschromatograph with electron capture detection (GC-ECD) for PAN analysis. A keystone in its design was the pre-concentration of PAN by capillary columns, which were used as up- and downdraft reservoirs for the HREA application and enabled simultaneous sampling at two measurement heights, which is an important feature for the MBR method. A major challenge of the design was the resolution of small PAN mixing ratio differences by the analytical unit, which required an optimum choice of operational settings and a detailed error analysis. The PAN flux measurement system was first applied on a nutrient-poor natural grassland site at the premises of the Mainz-Finthen Airport, Germany, in the period from August to September 2011. The application of the MBR method yielded average daytime PAN fluxes of -0.07 nmol/m2/s with a random flux error of ±0.03 nmol/m2/s, which was mainly attributed to the small PAN mixing ratio differences. Due to both a higher surface resistance and larger uncertainties of the PAN analysis, at the natural grassland site no significant PAN fluxes could be resolved with the HREA method. The thesis demonstrates that above low vegetation, like the studied grassland, the MBR method is in most cases more suitable than the REA technique, while the REA is preferably used above high vegetation such as forest canopies. Due to large errors of the HREA application, this thesis presents a detailed analysis of the effects of an imprecise sampling of up- and downdraft events for REA applications. Especially, the implication of a long inlet tube may introduce a significant lag time error and high frequency attenuation effects. The simulation of REA fluxes of several scalar quantities revealed that REA fluxes might be generally underestimated from less than 5% to about 50% for typically observed lag time errors. The observed effect of high frequency attenuation on REA fluxes is of similar magnitude, ranging between less than 5% to about 30% for typical filter strengths. For both effects a function of the flux loss with the so-called eddy reversal frequency was found, which could be used to correct HREA fluxes at the investigated grassland site or identify their uncertainties. Furthermore, in the Appendix a detailed procedure is provided how to minimize, evaluate and correct lag time and high frequency effects for future REA setups. In the last part of the thesis, the impact of PAN deposition at the nutrient-poor natural grassland site is investigated. PAN deposition fluxes obtained with the MBR method were partitioned into stomatal and non-stomatal deposition pathways. A significant non-stomatal conductance was found, which was of similar magnitude than the non-stomatal conductance for O3, which would imply that currently applied deposition models may significantly underestimate the deposition of PAN. The retrieved information on the stomatal and non-stomatal conductances was used to model the PAN deposition over a 3.5 months period (summer to early autumn). Since the measurement site was situated on the edge of an urban and industrialized region, this allowed the investigation of the influence of local air pollution on the PAN deposition. Although PAN deposition was about twice as large during polluted periods, the results suggest that PAN deposition does not play a critical role as a nitrogen source to the investigated nutrient-poor grassland ecosystem, during both polluted and unpolluted conditions. Instead, PAN deposition was found to contribute to 20-30% of the PAN removal from the atmosphere during daytime. In higher latitudes or during winter, when thermochemical decomposition of PAN is low, PAN deposition is likely to be the predominant sink.

Level thesis:degree_level

thesis.doctoral

Grantor dc:publisher

Universität Bayreuth

Year

2014

Author dc:creator

• Moravek, Alexander

Contributors dc:contributor

• Foken, Thomas

Repository record source_url

https://epub.uni-bayreuth.de/id/eprint/1688/

OAI identifier oai:identifier

oai:epub.uni-bayreuth.de:1688