An Air Quality Assessment of the European Nested Grid of the GEOS-Chem Model and the Influence of the Agricultural Sector

Air pollutants have a serious influence on public health. The ability of a chemical transport model (CTM) to capture pollutant concentrations is paramount to understanding air quality issues. For the first time the European nested (0.25o x 0.3125o) grid version of the GEOS-Chem model of chemistry and transport has been compared to European wide observations. Comparison between the model and the European Airbase monitoring network showed that the model is capturing Nitrogen Dioxide (NO2) seasonality albeit with a negative bias. The model was found to not be capturing the observed rise in NO2 rush hour concentrations across Europe sufficiently, suggesting that a re-evaluation of the transport NOx source in Europe is needed. For ethane and propane the model captured seasonality in concentrations but had a negative bias which was found to be likely from a missing source. This source is potentially gas leaks from the natural gas network not being accounted for in the model. The model captures ozone European seasonal concentrations well, with a small positive (between 2-15 nmol mol-1 across regions), bias that is worse during the night. The afternoon peak in ozone is well represented by the model in Europe, which is important for air quality applications. For Particulate Matter with diameter of less than 2.5 µm (PM2.5), the model captures seasonality well, with a small positive bias (5-11 μg m-3 across regions), that is exacerbated during periods of elevated concentration. Comparing modelled PM2.5 composition against the observational UK MARGA network, showed the model performed reasonably well and confirmed the modelled overestimation in PM2.5 was likely due to an overestimation of nitrates and sulphates in the model, which is likely due to uncertainties in the ammonia emissions inventory. The ability of the GEOS-Chem model to capture PM2.5 short term events was assessed. The model captured the timing of these high concentrations. There was differing reasons as to why these PM events occured. Both vehicular emissions (NOx) and agricultural emissions (NH3) played a role. From this we conclude that the March and April 2014 events were nitrate driven, highlighting the agricultural sector was a key component in exacerbating PM levels during the pollution event. Removing all agricultural emissions from the model has a profound impact on the european PM2.5 with the domain-mean surface PM2.5 concentration, with a fractional difference typically between 0.35 and 0.65 when emissions are switched off. For some regions (e.g. The Po Valley, central Europe, Turkey, and parts of north western Europe) there are large decreases in the number of days PM2.5 exceeds the WHO 25 μg m-3 limit (upto ~225 days, ~125 days, ~110 days and ~90 days respectively). Policies that target the reduction in agricultural ammonia emissions are likely to result in a significant reduction in the concentration of surface PM2.5 over Europe. Overall the model shows many positive traits, particularly capturing in ozone and PM2.5. However, it requires more extensive evaluation and comparison to observations before there can be confidence in its suitability for use in develop by European wide air quality policy and science.