Ensemble forecasts of air quality in eastern China – Part 1: Model description and implementation of the MarcoPolo–Panda prediction system, version 1

Abstract

Abstract. An operational multi-model forecasting system for air quality including nine different chemical transport models has been developed and provides daily forecasts of ozone, nitrogen oxides, and particulate matter for the 37 largest urban areas of China (population higher than 3 million in 2010). These individual forecasts as well as the mean and median concentrations for the next 3 days are displayed on a publicly accessible website (http://www.marcopolo-panda.eu, last access: 7 December 2018). The paper describes the forecasting system and shows some selected illustrative examples of air quality predictions. It presents an intercomparison of the different forecasts performed during a given period of time (1–15 March 2017) and highlights recurrent differences between the model output as well as systematic biases that appear in the median concentration values. Pathways to improve the forecasts by the multi-model system are suggested.

Prediction of source contributions to urban background PM<sub>10</sub> concentrations in European cities: a case study for an episode in December 2016 using EMEP/MSC-W rv4.15 and LOTOS-EUROS v2.0 – Part 1: The country contributions

“…Thus, the scenario approach is more appropriate for the calculation of the source contribution of the primary PM components than for nonlinear species such as the secondary components (e.g. Burr and Zhang, 2011;Thunis et al, 2019). LOTOS-EUROS traces the origin of air pollutants throughout a simulation using a labelling approach.…”

“…EMEP Status Report, 2018), and the LOTOS-EUROS chemistry transport model (Manders et al, 2017) has also been used in several source apportionment studies over Europe, especially for PM (Hendriks et al, , 2016Schaap et al, 2013). Both models are involved in the operational air quality analysis and forecasting for Europe in the CAMS regional ensemble system (Marécal et al, 2015) and for China (Brasseur et al, 2019). For the simplicity of the reading, the EMEP/MSC-W model is hereafter referred to as the EMEP model.…”

Response to short comment by Dr. P. Nowack

“…We use the random forest regression (RFR) algorithm (Breiman, 2001) to emulate the integration of atmospheric chemistry. Figure 2 shows a schematic of RFR.…”

“…The RFR algorithm is less prone to over-fitting and produces predictions that are more stable than a single decision tree (Breiman, 2001). Random forests are widely used since they are relatively simple to apply, suitable for both classification and regression problems, do not require data transformation, and are less susceptible to irrelevant or highly correlated input features.…”

Sensitivity of <i>CHIMERE</i> to changes in model resolution and chemistry over the northwestern Iberian Peninsula

“…In this context, most of the aforementioned conclusions for CHIMERE (the model applied here) have been drawn for the Île de France region, which is densely populated, relatively flat and not directly influenced by sea-salt emissions. The model has been applied for a number of other regions, but the map is still incomplete and sensitivity testing is not the main focus of the corresponding studies (Mazzeo et al, 2018;Menut et al, 2018;Monteiro et al, 2018;Brasseur et al, 2019;Deroubaix et al, 2019). This is where the present study comes into play: A series of sensitivity test has been run with CHIMERE over the northwestern Iberian Peninsula, a region characterized by a complex coastline, forested mountain terrain and the advection of sea-salt from the surrounding Atlantic Ocean.…”