A simulation model of the epidemiology of barley yellow dwarf virus in winter sown cereals and its application to forecasting.

Published online
15 May 1993
Content type
Journal article
Journal title
Journal of Applied Ecology
DOI
10.2307/2404510

Author(s)
Kendall, D. A. & Brain, P. & Chinn, N. E.

Publication language
English
Location
UK

Abstract

A computer model is described which simulates the spread of barley yellow dwarf luteovirus (BYDV) by aphids in winter cereals in SW England. The model has as a theoretical base the dynamic interaction between 2 temp. dependent rate processes: (i) the frequency of virus transmission to uninfected plants by infected aphids, and (ii) the frequency of virus acquisition by uninfected aphids from infected plants. A distinction is made between virus spread into crops by alate migrant vectors (primary transmission), and more localized spread within crops mainly by apterous vectors (secondary transmission). Both types of spread may occur concurrently. Probabilities of multiple infection, and effects of this on the rate processes in the model, are described by a generalized frequency distribution. Vector dispersal and the latent periods, following transmission and acquisition of virus, are represented as constant functions of thermal time. Simulations used field data of plant and aphid populations sampled every few weeks in crops of winter barley and wheat at Long Ashton near Bristol during 1978-89. The infectivity of autumn migrant aphids caught each year in a suction trap was measured by feeding transmission tests. Daily max-min screen temp. used to estimate thermal time were obtained from records at Long Ashton. Crop infections were assessed from visible symptoms or by ELISA. The model was successfully validated for 61 cereal crops with 0-80% of the crop area affected. There was no significant divergence between field observations and incidence of BYDV as predicted by the model (barley, r = 0.98; wheat, r = 0.94). Curves of epidemic developments obtained by measuring virus incidence at frequent intervals during crop growth were usually similar to those generated by the model. Prospects for practical and more reliable forecasting of BYDV epidemics, and of the need for control measures, are discussed.

Key words