Distribution and amounts of nitrous and nitric oxide emissions from British soils
This study establishes an empirical predictive model of N20 and NO emissions for Great Britain based on multivariate regression analysis of field measurement data from several studies in European countries and the USA for which the results have been published in the last 18 years. The significance of studying the emissions of these gases is due to the role of N20 as a greenhouse effect gas and NO participation in reactions with ozone. Soils are known to be an important source of N2O and also contribute significant amounts of NO into the atmosphere. Knowledge of N20 and NO emissions from soils at a national scale is important due to the signed international agreements which oblige Great Britain to produce inventories of greenhouse effect gases and monitor the emissions of NOx gases. The field studies observed the relationships between the emissions and their controlling factors and on the basis of those relationships, national modelling approaches to predicting the amounts of emissions have been defined. Due to the highly variable nature of emissions, more than one empirical model was developed for each of the gases. The relationships defined in the analysis were later applied to estimate N20 and NO emissions from British soils with an application of input parameter data of the established controlling factors in the framework of Arclnfo GRID.Data for N fertiliser input, soil moisture and temperature were not readily available and therefore had to be estimated with existing data. Soil moisture was predicted with the SPACTeach model based on the monthly precipitation sums obtained from the Climate LINK data set. This data source also provided monthly air temperature data used to model soil temperature with the theory of heat flux. N input was estimated as a sum of mineral and organic N fertiliser inputs from agriculture and atmospheric N deposition. The former was estimated from the recommended values according to spatial distribution of land use data provided by the Agricultural Census. The latter was based on the modelled N atmospheric deposition provided by the Review Group on Acid Rain (RGAR). Information on the extent of seminatural land was obtained from the Land Cover Map of Great Britain based on satellite data. The data sets were characterised by varied spatial resolution and were brought to a universal 5 km grid resolution prior to modelling emission as this was the best assumed resolution at the national scale. The predicted total of N20 emissions from British soils ranged between 128 and 140 kt N yl, and NO between 7 and 66 kt N y- 1, depending on the applied model. The predicted totals of N20 are higher than the estimates based on the approach of Bouwman (1995) and Skiba et al., (1996) using N emission factors. The higher NO emission rates based explicitly on the N factor suggest that the other approaches underestimate their totals (Simpson et al., 1999). The lower NO predicted in this study was due to the limiting effect of soil moisture. The different results of the models presented here are the result of the improved modelling approach used in this study, which takes into account the climatic characteristics of soils in addition to N input.The validation of the established models against field measurements from selected studies in Scotland showed their limited accuracy in predicting N20 and NO emissions at field scales. This was expected due to great spatial and temporal variability of emissions and the restricted methods of field measurements. While mechanistic models are better designed to reflect the emission processes at small scales, at national scales N20 and NO emissions are better predicted with simple regression models. This is mainly the result of limited availability of input data for large scale studies.