The Potential Global Climate Suitability of Kiwifruit Bacterial Canker Disease (Pseudomonas syringae pv. actinidiae (Psa)) Using Three Modelling Approaches: CLIMEX, Maxent and Multimodel Framework

In recent years, outbreaks of kiwifruit bacterial canker (Pseudomonas syringae pv. actinidiae, Psa) have caused huge economic losses to two major global kiwifruit producers, Italy and New Zealand. To evaluate the potential global risk areas of Psa, three modelling methods (MaxEnt, CLIMEX and a Multi-Model Framework, including Support Vector Machine or SVM) were used. Current global occurrence data for Psa were collected from different sources. The long-term climate data were sourced from WorldClim and CliMond websites. The model results were combined into a consensus model to identify the hotspots. The consensus model highlighted the areas where two or three models agreed on climate suitability for Psa. All three models agreed with respect to the climate suitability of areas where Psa is currently present and identified novel areas where Psa has not established yet. The SVM model predicted large areas in Central Asia, Australia, and Europe as more highly suitable compared to MaxEnt and CLIMEX. Annual mean temperature and annual precipitation contributed most to the MaxEnt prediction. Both MaxEnt and CLIMEX showed the probability of Psa establishment increased above 5 degrees C and decreased above 20 degrees C. The annual precipitation response curve showed that excessive rain (>1200 mm/y) constrains Psa establishment. Our modelling results will provide useful information for Psa management by highlighting the climatically susceptible areas where Psa has not established, such as the USA, Iran, Denmark, Belgium and especially South Africa, where kiwifruit has been planted commercially in recent years.

Automated summary

In recent years, outbreaks of kiwifruit bacterial canker have caused economic losses to major kiwifruit producers. Three modelling methods were used to evaluate the potential global risk areas of the disease. The models identified current areas of presence as well as novel areas where the disease has not established yet. The results will be useful for disease management and highlight susceptible areas where the disease has not established.