The temporal and spatial variability of soil respiration in boreal forests : a case study of Norunda forest, Central Sweden

Abstract: Previous studies have shown that, contrary to expectations, the boreal forest at the Norunda research site in central Sweden is acting as a carbon source. Therefore, the aim of this study was to investigate the spatial and temporal variability of soil carbon dioxide (CO2) efflux (FSoil) in order to provide insights into its role for the net carbon balance of this forest. The investigation is divided into three parts: (1) a general description of the soils, (2) the analysis of the spatial variability and (3) the analysis of the temporal variability of FSoil. Soil samples were collected at three soil profiles and analysed for physical and chemical properties. Measurements of FSoil and auxiliary, environmental data were carried out at 90 plots in the forest. The impact of environmental factors on the spatial variability of FSoil were investigated using a multiple, linear regression. Based on time series of FSoil and net ecosystem exchange (NEE), which are continuously measured at the study site, monthly and annual means of CO2 fluxes were derived and the contribution of soil respiration to ecosystem respiration was calculated. Furthermore, the correlations between FSoil, soil temperature (TSoil) and gross primary production (GPP) were investigated using the wavelet analysis. Soil pH as well as total C and N stocks were in typical ranges for boreal forest soils. The analysis of the spatial variability of FSoil yielded a coefficient of variation of 46 %. Only soil temperature and moisture had a significant effect on FSoil, they explained about 27 % of its variance. Stone content of the soil, distance to and species of the closest three trees were not found to be significant. Annual means of FSoil were found to be very high and are about 4-5 times as large as mean, annual NEE. Furthermore, it was found that, on average, soil respiration contributes 70%-100% of ecosystem respiration. While 70% is in agreement with previous studies, a value of 100% appears to be unrealistic and potential sources of errors should be further investigated. The wavelet analysis revealed significant correlations of FSoil, TSoil and GPP at the daily and yearly cycle, as expected. Additional correlations with periods of several days to weeks were found and are likely related to weather phenomena. Soil moisture appeared to affect the dynamics between FSoil and GPP. An analysis of phase differences revealed a time lag of 1-5 days between GPP and FSoil, confirming findings of previous studies. The results of this study highlight the importance of FSoil for the carbon balance of Norunda forest. However, in order to reliably estimate the contribution of soil respiration to ecosystem respiration, further investigations are required to rule out potential errors, as the results indicate that the continuously measured FSoil values are comparatively high. Furthermore, the variability of FSoil at the spatial scale could only be partly explained. Future studies on drivers of FSoil, which include more potential predictor variables and apply a non-linear, multiple regression, might yield further valuable insights.