Research

Assimilation of observational data

Example of modelled AOD obtained with combined assimilation of ground observed PM from the EMEP database and MODIS AOD (project GEOmon)

Introduction
Present monitoring techniques -when independently used- are not sufficient to capture the variability of pollutant fields in space and time. Both at the national and international level there is a large interest for improved air quality forecasts and analyses of ozone, NO2 and aerosols.
Utilizing observations in combination with models currently seems to be a promising chance to achieve these improved quality of the analyses and forecasts at different spatial scales (from the planetary scale to the local city, or even street scale) and time scales.
This integrated air quality modelling, where ground-based observations (in-situ and surface remote sensing) and satellite retrieved data are combined with chemical transport modelling, could lead to a large step in our ability of air quality monitoring and prediction.
Data-assimilation over Europe
Correlation between AERONET AOD and AOD as a result of data-assimilation with MODIS AOD (project GEOmon)

Applications of data assimilation arise in many fields of geosciences, perhaps most importantly in weather forecasting. A straight forward step is to exploit this technique for the integration of observations and modelling for improved analyses and forecasting of air quality.
For aerosols, the added value of assimilation of ground-based PM observations, of satellite derived aerosol optical depth (AOD), and of both data sources together, in the LOTOS-EUROS model is investigated. Part of this work is conducted under the flag of GEOmon project within the 6th Framework Programme of the EU contributing to The Global Earth Observation System of Systems (GEOSS). For ozone, the method of combined assimilation is presently succesfully applied for the operational forecasting of smog in the Netherlands, a result of the Dutch project SmogProg. Additionally, improved analysed maps for regulatory pollutants and air quality forecasts for the European area from combined assimilation model runs with LOTOS-EUROS are or will be produced in two projects falling under the GMES programme, which is funded by the 7th Framework Programme of the EU (MACC, PasoDoble). Here both groundbased ozone as well as tropospheric NO2 columns from the OMI satellite will be assimilated to improve the models performance. The assimilation of tropospheric NO2 columns will also be used to infer NO2 emission trends over Europe within the 7th Framework Programme project EnerGEO.
Assimilation method
To feed the LOTOS-EUROS model with observations, we make use of an ensemble Kalman filter (EnKF) data-assimilation technique. The EnKF method is applied by generating random noise to certain model input parameters (usually the emissions, deposition velocities and or top boundary conditions) to define the ensemble and therewith the model uncertainty. The method is an example of an active data assimilation approach, where forecast and analysis steps follow each other sequentially in time and use only information from the past. At each time step, the assimilation system delivers an ensemble of estimates of the spatial distribution of a particular pollutant with a corresponding uncertainty (taking into account the uncertainties in the measurements as well as the model).