Abstract. A reliable identification of Desert Dust (DD) episodes over North-central Spain is carried out based on AErosol RObotic NETwork (AERONET) columnar aerosol sun-photometer (aerosol optical depth, AOD, and Ångström exponent, α) and European Monitoring and Evaluation Programme (EMEP) surface particulate mass concentration (PMx, x = 10, 2.5, and 2.5–10 µm) as main core data. The impact of DD on background aerosol conditions is detectable by means of aerosol load thresholds and complementary information provided by HYSPLIT (Hybrid Single Particle Lagrangian Integrated Trajectory Model) air mass back-trajectories, MODIS (Moderate Resolution Imaging Spectroradiometer) images, forecasting aerosol models, and synoptic maps, which had been carefully reviewed by a human observer for each day included in the DD inventory. This identification method allows the detection of low and moderate DD intrusions and also mixtures of mineral dust with other aerosol types by means of the analysis of α. During the period studied (2003–2014), a total of 152 DD episodes composed of 419 days are identified. Overall, this means ~ 13 episodes and ~ 35 days per year with DD intrusion, representing 9.6 % days/year. During the identified DD outbreaks, 19 daily exceedances over 50 µg m−3 are reported at the surface. The occurrence of DD event days along the year peaks in March and June with a marked minimum in April and lowest occurrence in winter. A large inter-annual variability is observed showing a statistically significant temporal decreasing trend of ~3 days/year. As a key point, the DD impact on the aerosol climatology is addressed by evaluating the DD contribution to AOD, PM10, PM2.5, and PM2.5–10 obtaining mean values of 0.015 (11.5 %), 1.3 µg m−3 (11.8 %), 0.55 µg m−3 (8.5 %) and 0.79 µg m−3 (16.1 %), respectively. Almost similar annual cycles of DD contribution are obtained for AOD and PM10 with two maxima, one in summer (0.03 and 2.4 µg m−3 for AOD in June and PM10 in August, respectively) and another in March (0.02 for AOD and 2.2 µg m−3 for PM10), discrepancies occurring only in July and September. It is worth mentioning that the seasonal cycle of DD contribution to AOD does not follow the pattern of the total AOD (near bell shape), meanwhile both PM10 cycles (total and DD contribution) present more similar shapes between them, although a main discrepancy is observed in September. The inter-annual evolution of the DD contribution to AOD and PM10 has evidenced a progressive decrease. This decline in the levels of natural mineral dust aerosols can explain up to the 30 % of the total aerosol load decrease observed in the study area during the period 2003–2014. The relationship between columnar and surface DD contributions is evident with a correlation coefficient of 0.81 for the inter-annual averages. Finally, synoptic conditions during DD events are also analysed observing that the North African thermal low causes most of the events (~ 53 %). The results presented in this study highlight the relevance of the area studied since it can be considered as representative of the clean background in Western Mediterranean Basin where DD events have a high impact on aerosol load levels.
Supplementary material to "INVENTORY OF AFRICAN DESERT DUST EVENTS IN THE NORTH-CENTRAL IBERIAN PENINSULA IN 2003–2014 BASED ON SUNPHOTOMETER-AERONET AND PARTICULATE MASS-EMEP DATA"
