Modelling of Grass Pollen Interannual Variation in the UK: A Statistical Approach

Kurganskiy, Alexander ORCID: https://orcid.org/0000-0002-6588-9387, Creer, S., de Vere, N., Griffith, G., Osborne, N.J., Wheeler, B.W., McInnes, R.N., Clewlow, Y., Barber, A., Brennan, G.L., Hanlon, H.M., Hegarty, M., Potter, C., Rowney, F., Adams-Groom, Beverley ORCID: https://orcid.org/0000-0002-1097-8876, Petch, Geoffrey, Pashley, C.H., Satchwell, J. and Skjøth, C. ORCID: https://orcid.org/0000-0001-5992-9568 (2019) Modelling of Grass Pollen Interannual Variation in the UK: A Statistical Approach. In: European Aerosol Conference, 25th - 30th August 2019, Gothenburg, Sweden. (Unpublished)

	Text Kurganskiy_et_al_EAC2019_250_words_abstract_final.docx Restricted to Repository staff only Download (15kB) | Request a copy
	Text Kurganskiy_et_al_EAC2019_poster_a0_final.pdf Restricted to Repository staff only Download (3MB) | Request a copy

Abstract

Grass pollen is the most allergenic pollen type in the UK and up to 30% of the UK population are sensitized to grass pollen. Estimating the grass pollen season strength and interannual variation is a key task in aerobiological studies. The season strength is quantified using the Seasonal Pollen Integral (SPIn). The SPIn is the integral over time of daily pollen concentration and an important component in forecast models. Commonly, the SPIn interannual variation is modelled applying a statistical approach by analysing 10-20 years of pollen data from one observational site and meteorological observation data for the specific region. The statistical approach employed in this study is devoted to build a geostatistical regression model that goes beyond traditional approaches. Grass pollen SPIn and pre-seasonal air temperature observations are used as input data for 14 pollen observation sites covering all of the UK, totalling 176 years of data. The modelled SPIn has been compared with observations by means of statistical analysis. The results show that the model explains 63% of the observed variance found throughout UK and has a little bias (FB = -0.03) at the selected sites. The study also suggests that the SPIn, hence annual pollen exposure, is largely governed by pre-seasonal meteorological conditions. The geostatistical regression model therefore has the potential to be implemented in numerical models to simulate grass pollen concentrations within a forecasting environment.

Actions (login required)

View Item