ECN publication
Methode ontwikkeling voor verbetering van de verificatie van niet-CO2 broeikasgasemissies
Published by: Publication date:
ECN Biomass, Coal and Environmental Research 1-12-2004
ECN report number: Document type:
ECN-C--04-027 ECN publication
Number of pages: Full text:
55 Download PDF  

The main purpose of the project was the improvement of the accuracyof emission verification strategies for non-CO2 greenhouse gases on the European and regional (countrywise) scale. Two Lagrangian type of air pollution dispersion models have been adapted and were utilized in order to arrive at accurate description of process fluxes at the regional scale. The most important improvements in the ECN COMET model were: better and higher resolu-tion 3D trajectory data and a better parametrisation of the mixing layer height. Furthermore measured vertical concentration gradients measured at the tall tower at Cabauw (NL) were used compared to the previous use of data for one measurement height only. The new version of COMET produces a much better comparison between concentration measurements and the model values, especially for the nighttime accumulation values. The linear regression of hourly data for the whole year of 2002 of methane between observations and the model delivers a cor-relation coefficient (r2) of 0.69. This means that the explained variation between measurements and model values has improved with almost a factor of two. From forward and inverse calculation we find that mixed layer averaged concentrations of methane at Cabauw are mainly determined by sources within The Netherlands. In the study pe-riod the contribution of The Netherlands to the methane in air at Cabauw in excess to the back-ground is 65-80%. The total emission of the 6 areas of The Netherlands is estimated by the model inversions with a rather stable emission rate of 70-80% of the value of the estimate of the emission inventory METDAT, i.e. 1020 kton/yr for the base year 1995. The uncertainty analysis using a Monte Carlo error simulation technique indicates a rather high uncertainty of the individual emissions of the six separate emission areas in The Netherlands. The uncertainty is higher for the whole 2002 year period than for a shorter six week period in September/October 2002. After updating the emissions with the optimised values as calculated with the inverse COMET method, the correlation coefficient between observations and model values does not improve significantly. The Root Mean Square Error however is decreased substantially to about half the value obtained when using the uncorrected emissions. Using the optimized emission field also reduces the night-time overestimation of the modeled concentrations. The present results do not allow to discern the spatial distribution of the different source catego-ries over The Netherlands. This can be attributed to several reasons like the error introduced by discretisation of the source areas, numerical uncertainties in the (big) solution space and the too high background concentrations used here (TM3, based on 1993 data).

Back to List