Modelling the Spatial Distribution of Culicoides imicola: Climatic versus Remote Sensing Data

Remote Sensing

Guy Hendrickx¹  Els Ducheyne¹  Bernard De Baets²  Jan Peters²  Jasper Van Doninck³  Niko E.C. Verhoest³ 

[1] Avia-GIS, Risschotlei 33, 2980 Zoersel, Belgium;Department of Mathematical Modelling, Statistics and Bioinformatics, Ghent University, Coupure links 653, 9000 Ghent, Belgium;Laboratory of Hydrology and Water Management, Ghent University, Coupure links 653, 9000 Ghent, Belgium;

关键词: species distribution modelling;    bluetongue;    MODIS;    WorldClim;    random forests;    variable importance;   

DOI  :  10.3390/rs6076604

来源: DOAJ

【 摘 要 】

Culicoides imicola is the main vector of the bluetongue virusin the Mediterranean Basin. Spatial distribution models for this speciestraditionally employ either climatic data or remotely sensed data, or acombination of both. Until now, however, no studies compared theaccuracies of C. imicola distribution models based on climaticversus remote sensing data, even though remotely sensed datasets mayoffer advantages over climatic datasets with respect to spatial andtemporal resolution. This study performs such an analysis for datasetsover the peninsula of Calabria, Italy. Spatial distribution modellingbased on climatic data using the random forests machine learningtechnique resulted in a percentage of correctly classified C. imicola trappingsites of nearly 88%, thereby outperforming the linear discriminantanalysis and logistic regression modelling techniques. When replacingclimatic data by remote sensing data, random forests modellingaccuracies decreased only slightly. Assessment of the differentvariables’ importance showed that precipitation during late spring wasthe most important amongst 48 climatic variables. The dominant remotelysensed variables could be linked to climatic variables. Notwithstandingthe slight decrease in predictive performance in this study, remotelysensed datasets could be preferred over climatic datasets for themodelling of C. imicola. Unlike climatic observations, remotesensing provides an equally high spatial resolution globally.Additionally, its high temporal resolution allows for investigatingchanges in species’ presence and changing environment.

【 授权许可】

Unknown