By Truong Ngoc Phuong
Abstract:
The INTEGRATOR model (Integrated Nitrogen Tool across Europe for GReen House Gas and Ammonia Targeted to Operational Responses) is the integrated, GIS-based, simple multi-sector, multi-component modelling tool for estimating GHGs emission at European scale. It will be used for tracking changes over time for assessment of the success of policy measures and the impact of land use changes on the overall European N and GHG budgets. It comprises of multiple sub models to predict GHGs emission including terrestrial nitrous oxide (N2O) which is the target of uncertainty analysis in this study. To calculate terrestrial N2O emission from natural and agricultural, the INTEGRATOR model uses two different models which contain different sources of error. Uncertainty about models’ inputs (soil pH and soil organic carbon), model parameters (emission factors and coefficients in relationship describing N2O exchange rate from soil ) and model structure (random error in empirical model) are quantified based on defining characteristics of their probability distribution. Uncertainty propagation to the output is quantified following Monte Carlo approach. Analysis of spatial aggregation effect on uncertainty in estimate of N2O emission are carried out by aggregating the outputs from point scale of 10 by 10 kilometres spatial interval to national scale. The main results of this study indicates that: (1)The developed statistical models are sufficient to quantifying uncertainty about model inputs, model parameters and model structure; (2) Uncertainty about N2O estimate is expressed by the standard deviation which varies from very low at 0.007 kg N/ha/yr to around 16.58 kg N/ha/yr over all examined locations; (3) Error in emission factors and empirical model are the most importance sources of error in calculaing terrestrial N2O emission by the INTEGRATOR model; (4) Spatial aggregation from point scale to national scale do reduce much of uncertainty in the estimated terrestrial N2O emission and degree of reduction is proportional to extent of area aggregated to.