Page - 123 - in Intelligent Environments 2019 - Workshop Proceedings of the 15th International Conference on Intelligent Environments

per sites in several cases. The sites are characterized by geographic coordinates (Lon- gitude andLatitude). These traitswere used as a response or dependent variable in the modeling. Figure1. ICARDAbarleyaccessions Table1. Characterization traitsused forbarleymodeling Trait Accessionsevaluated Uniquesite evaluation trait values Days toheading 15027 3779 early, late Days tomaturity 15012 3775 early, late Kernelweight 5413 1881 low,high Productive tilleringcapacity 10286 2844 low,high Kernel covering 18220 4562 naked, covered Kernel rownumber 20256 4610 six-rowed, two-rowed Growthclass 18376 3932 winter, spring Yellowrust 1683 756 resistant, susceptible 2.2. Predictors:WorldClimandENVIREMData In themodeling,weusedenvironmentaldata fromWorldClim1 andEnvirem2 databases aspredictors. WordClimData WorldClimisanopenaccessdatabaseprovidingglobalclimatic layers describingpast climatic profiles of collection sites and intended for spatialmodelingor mapping. It includesaverage,monthlyminimumandmaximumtemperatures,precipita- tionandbioclimaticvariables [11]. 1https://worldclim.org/ 2https://envirem.github.io/ Z.Azoughetal. /PredictiveCharacterizationof ICARDAGenebankBarleyAccessions 123