Nitrous oxide emission from soil : On extrapolation from soil environmental factors

Abstract: Nitrous oxide(N20) is a natural component of the atmosphere which contributes both to the greenhouse effect and to the destruction of stratospheric ozone. National as well as global inventories indicate thatboth natural and anthropogenic emissions from terrestrial ecosystems constitute a large fraction of the annual loading of N20 on the atmosphere. However, estimates of the magnitude of emissions from different ecosystems are based on rather poor information on annual rates and the spatial distribution in ecosystems and landscape elements.This study (including 5 papers) focuses on the possibility to improve the estimates of N20 emissions from soil, by relating them to soil environmental factors. Also, the use of remote sensing of soil thermalinfrared emission, as a measure of one environmental factor (soil moisture) is evaluated.More specifically, momentary N20 emission and denitrification activity as functions of soil moisture, soil organic C and N, water-soluble C, pH, soil N03-, NO£ and NH/ contents and soil temperature were studied. In two studies, soils, with different clay contents were included. N20 emissionwas measured with closed chambers and denitrification rates in cores or chambers supplied with acetylene. Soil factors were either measured in soil from cores used for denitrification activity estimations or in separate soil samples. Relationships were analyzed by traditional regression analyses and partial least squares analysis of latent variables (PLS). PLS analyses are not sensitive to interrelations between the independent variables. A geometric interpretation (plot) of the result of the regression is possible, which gives excellent opportunities to find clusterings of the data. This method is particularly valuable in cases with few observations.Results from traditional regression and PLS analyses were in good agreement. In two cases clusterings of the data were observed (and data was also split and analysed after groupings). Generally, a minor part of the variation of the N20 emission and the denitrification rated could be explained by soil factors. When clay soils (<o40% clay) were analyzed separately, 65% (R2) of the variation of the N20 emission rates could be explained by soil NH4+ content and soil moisture. In all other cases, less than 41% of the variation ofN20 emission and denitrification rates could be explained.In conclusion: there are limited possibilities to obtain better estimates of large-scale and long-term N20 emission by relating momentary N20 emission rates to soil environmental factors. Several reasons for the poor relationships obtained are suggested, e.g. discrepancies between measurements of soil factors in bulk soil and zones/sites of microbial activity or omission of important soil factors in the analyses. In the study of remote sensing of soil thennal infrared emission as a measure of soil moisture, it was concluded that in the 0-10 cm soil layer it could only be used for intermediate soil moisture levels. It can be used for a wider moisture range in the upper few centimeters, which may still be of value in studies ofN20 emission and denitrification.Finally, an alternative approach to arriving at estimates of large-scaleand long-term N20 emissions, focusing on ecosystems and landscape elements is suggested.