Temporal Geostatistical Analyses of N₂O Fluxes from Differently Treated Soils

Abstract

Both CO₂ and N₂O fluxes from soils can contribute in a relatively important way to the greenhouse effect so the study of these variables is highly relevant. The N₂ O flux from soil presents high spatial and temporal variability. Application of geostatistical techniques to daily data of N₂ O fluxes may reveal underlying temporal structures in this variable. The aim of this study was to analyse patterns of temporal dependence in soil N₂ O fluxes. The experimental design consisted of four 100 m x 150 m plots at the Elora Research Station (Ontario, Canada). The data set corresponded to the period from January 2001 to December 2004. An exponential model was fitted to 15 out of 16 series. The models fitted to the experimental variograms consisted of two structures; a nugget effect and an exponential or spherical model depending on the analysed data series. The nugget effect ranged from 0 to 70% of the sill depending on the plot and year analysed. From the models fitted to the experimental variograms a temporal structure was identified, which ranged between 9 and 85 days, depending on the year and plot analysed. Conditional simulation was applied in order to estimate N₂ O fluxes for days with missing data, proving to be an appropriate tool for achieving this purpose.