Totals and variability of the emissions of CR and CAMSIRAIn the paper, we focus on 7 regions (Europe, North America, East Asia, South America, North Africa Tropical Africa and Maritime South East Asia) as well as the tropics, the Arctic and the Antarctic.The areas are shown in Figure S1 and the coordinates of the bounding boxes are given in Table 3 of the paper.Table S1 list the annual mean, maximum and minimum emissions for different source regions for CO, NO, isoprene (C5H8,) and all aerosol components (desert dust, sea salt, black carbon, organic matter and SO2 tracer).An estimated linear trend, if significant with 95% confidence interval, is also listed.Time series of the anthropogenic and biomass burning emissions are shown in Figure S2.The anthropogenic MACCity emissions for CO and NO decrease over Europe and North America in the range of 1 to 5% per year, whereas they increase over South East Asia by a similar amount.The global trend is almost zero for CO and slightly positive for NO.The observed emissions from biomass burning as well as modelled emission of biogenic VOC, desert dust and sea salt depend strongly on the inter-annual variability of the meteorological situation but changes in land use also play a role for long term trends.Besides a large year-to-year variability, the biomass burning emissions for CO show no significant trend over the whole 2003-2015 period.However, a significant reduction occurred until 2013 when global CO biomass burning emissions decreased from 423Tg in 2003 to 288 Tg in 2013.The decreasing trend was most pronounced in South America (-7.5%/yr)where it occurred over the whole period.
[1] A near real-time system for assimilation and forecasts of aerosols, greenhouse and trace gases, extending the ECMWF Integrated Forecasting System (IFS), has been developed in the framework of the Global and regional Earth-system Monitoring using Satellite and in-situ data (GEMS) project. The GEMS aerosol modeling system is novel as it is the first aerosol model fully coupled to a numerical weather prediction model with data assimilation. A reanalysis of the period 2003–2009 has been carried out with the same system. During its development phase, the aerosol system was first run for the time period January 2003 to December 2004 and included sea salt, desert dust, organic matter, black carbon, and sulfate aerosols. In the analysis, Moderate Resolution Imaging Spectroradiometer (MODIS) total aerosol optical depth (AOD) at 550 nm over ocean and land (except over bright surfaces) was assimilated. This work evaluates the performance of the aerosol system by means of case studies. The case studies include (1) the summer heat wave in Europe in August 2003, characterized by forest fire aerosol and conditions of high temperatures and stagnation, favoring photochemistry and secondary aerosol formation, (2) a large Saharan dust event in March 2004, and (3) periods of high and low sea salt aerosol production. During the heat wave period in 2003, the linear correlation coefficients between modeled and observed AOD (550 nm) and between modeled and observed PM2.5 mass concentrations are 0.82 and 0.71, respectively, for all investigated sites together. The AOD is slightly and the PM2.5 mass concentration is clearly overestimated by the aerosol model during this period. The simulated sulfate mass concentration is significantly correlated with observations but is distinctly overestimated. The horizontal and vertical locations of the main features of the aerosol distribution during the Saharan dust outbreak are generally well captured, as well as the timing of the AOD peaks. The aerosol model simulates winter sea salt AOD reasonably well, however, showing a general overestimation. Summer sea salt events show a better agreement. Overall, the assimilation of MODIS AOD data improves the subsequent aerosol predictions when compared with observations, in particular concerning the correlation and AOD peak values. The assimilation is less effective in correcting a positive (PM2.5, sulfate mass concentration, Angström exponent) or negative (desert dust plume AOD) model bias.
Abstract. Land-atmosphere interactions govern the process of dust emission and transport. An accurate depiction of these physical processes within numerical weather prediction (NWP) models allows for better estimating the spatial and temporal distribution of the dust burden and the characterisation of source and recipient areas. In the presented study, the ECMWF-IFS (European Centre for Medium-Range Weather Forecast - Integrated Forecasting System) outputs are used to simulate two-month long periods in the spring and autumn of 2020, focusing on a case study in October. The ECMWF-IFS outputs are produced with and without assimilation of Aeolus quality-assured Rayleigh-clear and Mie-cloudy Horizontal Line of Sight (HLOS) wind profiles. The experiments have been performed over the broader Eastern Mediterranean and Middle East (EMME) region that is frequently subjected to dust transport, as it encompasses some of the most active erodible dust sources. Aerosol and dust-related model outputs (extinction coefficient, optical depth and concentrations) are qualitatively and quantitatively evaluated against ground- and satellite-based observations. Ground-based columnar and vertically resolved aerosol optical properties are acquired through AERONET sun photometers and PollyXT lidar, while near-surface concentrations are taken from EMEP. Satellite-derived vertical dust and columnar aerosol optical properties are acquired through LIVAS and MIDAS, respectively. Overall, in cases of either high or low aerosol loadings, the model predictive skill is improved when WRF simulations are initialised with IFS meteorological fields in which Aeolus wind profiles have been assimilated. The improvement varies in space and time, with the most significant impact observed for the autumn months in the study region. Comparison with observation datasets saw a remarkable improvement in columnar aerosol optical depths, vertically resolved dust mass concentrations and near-surface particulate concentrations in the assimilated run against the control run. Reductions of model biases, either positive or negative, and an increase in the correlation between simulated and observed values were achieved.
Abstract. In the framework of the World Meteorological Organisation's Sand and Dust Storm Warning Advisory and Assessment System, we evaluated the predictions of five state-of-the-art dust forecast models during an intense Saharan dust outbreak affecting western and northern Europe in April 2011. We assessed the capacity of the models to predict the evolution of the dust cloud with lead times of up to 72 h using observations of aerosol optical depth (AOD) from the AErosol RObotic NETwork (AERONET) and the Moderate Resolution Imaging Spectroradiometer (MODIS) and dust surface concentrations from a ground-based measurement network. In addition, the predicted vertical dust distribution was evaluated with vertical extinction profiles from the Cloud and Aerosol Lidar with Orthogonal Polarization (CALIOP). To assess the diversity in forecast capability among the models, the analysis was extended to wind field (both surface and profile), synoptic conditions, emissions and deposition fluxes. Models predict the onset and evolution of the AOD for all analysed lead times. On average, differences among the models are larger than differences among lead times for each individual model. In spite of large differences in emission and deposition, the models present comparable skill for AOD. In general, models are better in predicting AOD than near-surface dust concentration over the Iberian Peninsula. Models tend to underestimate the long-range transport towards northern Europe. Our analysis suggests that this is partly due to difficulties in simulating the vertical distribution dust and horizontal wind. Differences in the size distribution and wet scavenging efficiency may also account for model diversity in long-range transport.
'%3e%3cpath fill='%23012169' d='M0 0v30h60V0z'/%3e%3cpath stroke='%23fff' stroke-width='6' d='m0 0 60 30m0-30L0 30'/%3e%3cpath stroke='%23C8102E' stroke-width='4' d='m0 0 60 30m0-30L0 30' clip-path='url(%23b)'/%3e%3cpath stroke='%23fff' stroke-width='10' d='M30 0v30M0 15h60'/%3e%3cpath stroke='%23C8102E' stroke-width='6' d='M30 0v30M0 15h60'/%3e%3c/g%3e%3c/svg%3e)