Ecology of West Nile virus across four European countries: empirical modelling of the Culex pipiens abundance dynamics as a function of weather

Thomas A Groen,Dusan Petric,Marija Zgomba,Laurence Marrama,Romeo Bellini,Gregory L’Ambert,Dominique J Bicout,Alexandra Chaskopoulou

doi:10.1186/s13071-017-2484-y

Abstract

BackgroundCulex pipiens is the major vector of West Nile virus in Europe, and is causing frequent outbreaks throughout the southern part of the continent. Proper empirical modelling of the population dynamics of this species can help in understanding West Nile virus epidemiology, optimizing vector surveillance and mosquito control efforts. But modelling results may differ from place to place. In this study we look at which type of models and weather variables can be consistently used across different locations.MethodsWeekly mosquito trap collections from eight functional units located in France, Greece, Italy and Serbia for several years were combined. Additionally, rainfall, relative humidity and temperature were recorded. Correlations between lagged weather conditions and Cx. pipiens dynamics were analysed. Also seasonal autoregressive integrated moving-average (SARIMA) models were fitted to describe the temporal dynamics of Cx. pipiens and to check whether the weather variables could improve these models.ResultsCorrelations were strongest between mean temperatures at short time lags, followed by relative humidity, most likely due to collinearity. Precipitation alone had weak correlations and inconsistent patterns across sites. SARIMA models could also make reasonable predictions, especially when longer time series of Cx. pipiens observations are available.ConclusionsAverage temperature was a consistently good predictor across sites. When only short time series (~ < 4 years) of observations are available, average temperature can therefore be used to model Cx. pipiens dynamics. When longer time series (~ > 4 years) are available, SARIMAs can provide better statistical descriptions of Cx. pipiens dynamics, without the need for further weather variables. This suggests that density dependence is also an important determinant of Cx. pipiens dynamics.

Highlights

Culex pipiens is the major vector of West Nile virus in Europe, and is causing frequent outbreaks throughout the southern part of the continent
This study showed that temperature is the most consistent weather variable to predict Cx. pipiens dynamics in southern Europe, directly with simple linear models
We show that when long time series are available, (S)ARIMA models could be very useful

Summary

Introduction

Culex pipiens is the major vector of West Nile virus in Europe, and is causing frequent outbreaks throughout the southern part of the continent. Because of WNV outbreaks in several places in southern Europe, vector control programmes have been set-up or are under evaluation to control Cx. pipiens populations [3]. What can be considered as a high density of Cx. pipiens in one region can be considered as a low density in another region This can be influenced by a number of factors, including landscape characteristics and the weather conditions [5, 6]. Standardized statistical descriptions of vector dynamics can be useful to serve as a benchmark against which population reductions resulting from control programmes can be compared. This is, for example, how control efficacy in mathematically designed models is tested [7]

Methods

Results

Conclusion