Browsing by Subject "urban meteorology"

In order to reliably quantify the impact of cities to the climate change, carbon dioxide (CO2) fluxes from urban areas need to be accurately estimated. Currently, eddy covariance (EC) measurements of net ecosystem exchange (NEE) are among the most used methods for CO2 balance assessment at ecosystem scale. They cannot, however, directly separate gross primary production (GPP) from different sources of CO2 , but instead different tracers and methods need to be used for the separation. In this thesis, urban carbonyl sulfide (COS) fluxes are reported and used as a tracer for biogenic CO2 uptake. For estimating GPP, a leaf-relative uptake (LRU) of CO2 and COS during photo- synthesis has to be calculated. Three different methods for ecosystem scale LRU estimations are compared in order to find which performs the best at assessing urban GPP from EC measurements. LRUNEE method is based on averaging LRU obtained directly from an equation using EC measurements of CO2 and COS. LRUP AR uses parameterisation of photosynthetically available radiation (PAR) to estimate LRU, and LRUCAP uses additional environmental parameters and information of soil water content (SWC), water vapour pressure deficit (VPD) and PAR. EC measurements were conducted at SMEAR III station (ICOS ecosystem associate site) in Helsinki during summer 2022 using a quantum cascade laser gas analyser. Mole fractions of COS, CO2, CO, and water vapour were collected with 10 Hz frequency, together with three-dimensional wind speed measurements. 30-min fluxes were calculated and used for estimating LRU and GPP with the three methods above. Source area of the station was divided into urban, street, and vegetation sectors. To compare different LRU methods, EC measured GPP from vegetation sector was used because it had the smallest anthropogenic emissions. CO and COS fluxes were compared to estimate co-emissions from fuel combustion. Median COS flux of −25.5 pmol m−2 s−1 was measured, indicating a biogenic CO2 uptake in the study area. For simplicity, soil related COS fluxes were neglected. LRUPAR performed the best at estimating GPP, giving an average CO2 uptake of 13.9 μmol m−2 s−1 for the vegetation sector. Furthermore, peak median anthropogenic emissions of 9.1 μmol m−2 s−1 were deduced with LRUPAR, which agrees with earlier research. However, due to problems with instrumentation the data collection period was so short that distinction between LRUPAR and LRUCAP was not unambiguous. Clear is that setting a constant value for LRU leads to significant underestimation of GPP, and a misunderstanding of its behaviour in heterogeneous urban landscape. Further development of COS based GPP estimation should include a more careful instrumentation and revision of the parameterisation of LRUCAP method.

Growing population in cities increases the share of global greenhouse gas (GHG) emissions coming from urban areas. To understand the energy, water and GHG emission exchanges between urban surface and the atmosphere, modelling is a necessary tool. This is because measurements are not always available from all the different urban environments. In the case of carbon dioxide CO2 exchange, modelling is needed to provide new information on the different anthropogenic and biogenic components over various land uses. In this thesis, the aim was first to compare energy and CO2 fluxes from an urban land surface model called Surface Urban Energy and Water Balance Scheme (SUEWS) against measurements from suburban neighbourhood in Minneapolis, USA. The second aim was to study differences in the fluxes between years in the area. The model is parameterized with surface information about the study area, which is divided into two grids, residential and recreational area. The meterological forcing data are derived from ERA5. In the first part of the study, SUEWS is run in the area from June 2006 to April 2009, and the fluxes of latent QE and sensible QH heat and CO2 are compared against eddy covariance (EC) measurements conducted in the same area in the same time period. The diurnal cycles of CO2 show that the model is able to catch the daytime values well in every season for both study area grids, but night-time positive values are difficult especially for recreational area in autumn and winter. The model also underestimates the emissions in every season in the morning and evening rush hour peaks, which are caused by traffic. Overall, CO2 flux is simulated reasonably well. The model performs very well against QE measurements, but more poorly against QH. The second part of the study extended time period from January 1995 to April 2009 to analyze the long-term variation of fluxes. These were studied independently without the measurement comparisons. Annual cumulative sum of CO2 showed great variation between the years, and the highest value was emissions of 1135 gCm-2year-1 in 2001 and the lowest 600 gCm-2year-1 in 2005 from the residential area. Annual cumulative sums of QE did not show so much variation. The reason behind the differences between these two years was the great variation of photosynthesis. In 2001 air temperature restrained photosynthesis when surface conductance and its environmental factors were further studied. No statistical difference between the years 2001 and 2005 was though found.