Improvement in the accuracy of backtrajectories using WRF to identify pollen sources in southern Iberian Peninsula

Hernández-Ceballos, M.A.(1,*), Skjøth, C. A.(2), García-Mozo, H.(3), Bolívar, J.P.(1), Galán, C.(3)

(1) Department of Applied Physics, University of Huelva, Huelva, Spain.

(2) National Pollen and Aerobiology Research Unit, Institute of Science and the Environment, University of Worcester, United Kingdom.

(3) Department of Botany, Ecology and Plant Physiology, Agrifood Campus of International Excellence (CeiA3), University of Córdoba, Córdoba, Spain.

* Currently working at: European Commision, Joint Research Centre, Institute for Transuranium Elements, Nuclear Security Unit, Ispra, Italy.

Electronic Supplementary Material

Fig. S1 shows results from digital elevation model over the southern Iberian Peninsula at each of the four different spatial resolutions, 111 km from GDAS files (Fig. S1a), and 27, 9 and 3 km from WRF files (Fig. S1b-d). Large differences in the relief of the landscape are seen in these four figures. The geography of southern area is a clear example of how winds are conditioned by the landscape geography elements. This region is mainly characterized by the presence of Guadalquivir valley surrounded by two mountain chains to the north (Sierra Morena) and south (Beticas system). So, its well geographical representation plays a key role in the characterization of surface winds in this area. In this case, the presence of the valley across Cordoba province is only observed in the 9km and 3km resolution.

Fig S1. Digital elevation model of southern Iberian Peninsula at different resolutions, a) 111 km, b) 27 km, c) 9 km and d) 3 km

To confirm the pollen transport from the northern source areas to Cordoba city, the hourly pollen evolution at El Cabril site during 6-7 June is shown in Fig. S2. This figure displays how hourly pollen concentration on 6th June was marked by low values (< 200 pollen grains/m3) until 15:00-16:00 UTC, when a progressive and sharp increase in pollen concentrations was observed, reaching at 17:00-18:00 UTC the value of around 800 pollen grains/m3. After that, pollen concentration registered a decreased, reaching similar morning values. On 7th June, concentrations were lower than 400 pollen grains/m3. With this information, the relationship between daily pollen cycles at El Cabril and Córdoba is ensured, with a first peak at El Cabril which is reported five hours late at Cordoba city.

Fig S2. Hourly evolution of pollen concentrations at El Cabril and Cordoba sites on 6-7 June 2010.

Regarding the second peak, the comparison of daily pollen evolution during these days in several sites along the valley, as Huelva and Seville, reveals the presence of relative peaks on 5th and 6th June suggesting the transport along the Guadalquivir valley. Fig S3 shows the daily concentrations of olive pollen in Huelva, Seville and Cordoba during this period. It shows a peak in both emplacement on 5th June, previously to the large peak registered in Cordoba.

Fig S3. Daily evolution of pollen concentrations along the Guadalquivir valley in the sampling stations of Huelva, Seville and Cordoba on 3-9 June 2010.

Fig. S4 and S5 show the cross-section of the horizontal (u) component (isotachs) and specific humidity along the Guadalquivir valley axis ( ~320 km; from (37.10 N,-6.70 W) to (37.97 N,-4.42 W)) and along the northwest-souteast province axis (~120 km; from (38.10 N, -5.48W) to 37.67N, -4.17W)) at different hours during the sampling period. This information aims to help in the analysis of air mass movement that can justify the pollen concentrations at Cordoba city.

Fig S4. Cross section of specific humidity and u-wind isotach and vertical motion on 6 June 2010 at a) 18:00 UTC, b) 22:00 UTC and on 7 June 2010 at c) 03:00 UTC, d) 10:00 UTC along the Guadalquivir valley.

The dominance of westerly winds at the surface levels favours the progressive penetration of maritime air masses along the valley. The proof of this movement is observed in Fig.S4a, in which the maritime front is located 60–75 km far away at 18:00 UTC. In its movement, these air masses swept inland areas of the valley, reaching the surroundings of Cordoba city (Fig. S4b) at 22:00 UTC. The analysis also shows how their vertical extension decreased over time from 1000 m at 18:00 UTC to less than 300 m at 22:00 UTC.

At night, the wind behaviour tended to be undefined in the surrounding of Cordoba city in direction and intensity. The model forecasted the progressive establishment of easterly winds over Cordoba city, favouring the westward movement of the air masses. This movement coincided with the first pollen peak. After that, it observed the progressive arrival to the valley of northern and drier air masses, which swept the northern pollen sources previously (Fig. S5). So, the second pollen peak at Cordoba city can be associated with the sum of 1) pollen transport along the valley and 2) arrival of northern air masses. This presence is removed due to the establishment of westerly winds again along the valley, causing the eastward displacement of air masses over Cordoba city.

Fig S5 Cross section of specific humidity and u-wind isotach and vertical motion on 7 June 2010 at a) 03:00 UTC and b) 10:00 UTC along Cordoba city.

The analysis of the atmospheric boundary layer characteristics during this period plays an important role in the understanding the pollen behaviour (Fig. S6). The figure suggests that pollen transport is performed in a height lower than 100 m. This fact is in agreement with the air mass movements and also favoured the confinement of southern and northern air masses in lower layers of the valley. While high concentrations were reached, the ABL height is less than 100 m. Then the ABL started to increase with about a factor of 20, which decrease concentration dramatically. So, this is likely the cause to the decrease in the concentration. We have more or less the same (or even higher) number of pollen in the atmosphere, but the pollen is diluted in a much larger fraction of the atmosphere.

Fig S6 Hourly evolution of pollen concentrations and height of the Atmospheric Boundary Layer (ABL) at Cordoba site on 6-7 June 2010.