Investigating the Influence of Climate Change on West Nile Virus Occurrence in Ontario, Canada

Abstract

West Nile virus (WNV) first appeared in North America during the spring season of 2002, growing to be the most common mosquito borne disease in Ontario, and becoming the primary cause of viral encephalitis. Climate change is hypothesized to increase Ontarian contact with WNV as climatic variables have been studied to amplify bridge vector replication in infected mosquitos. To further investigate this relation, we developed 2 models—Logistic Regression and Decision Tree—to predict WNV occurrence in mosquitos by these factors in Ontario. The Decision Tree model reflected a higher overall accuracy of 100% in training and 93% in testing. Both the Decision Tree model and Logistic Regression model found positive correlations between factors of night temperature, night humidity, NDVI, and rainfall, to the occurrence of mosquito WNV output. While the Logistic Regression model itself, found the occurrence of mosquito WNV cases to have a negative correlation to the factors: wind speed, day humidity, and total precipitation. Our findings can provide useful insight on when to warn the public to take special measures against WNV disease.