Projecting spring wheat yield changes on the Canadian Prairies: effects of resolutions of a regional climate model and statistical processing

In addition to the uncertainty associated with crop models, climate scenarios are still a major source of uncertainty in projecting crop yield changes under climate change. Regional climate models (RCMs) are used as tools for dynamic downscaling of climate scenarios from global climate models (GCMs) to regional scales for climate change impact studies. It is known that running an RCM is more expensive and time consuming at a higher resolution than at a lower resolution. Therefore, it is interesting to investigate how resolutions of an RCM and statistical processing of RCM data may result in differences in projected crop yield changes. We used the decision support system for agrotechnology transfer (DSSAT)–CERES-Wheat model to simulate yield changes of spring wheat at 13 locations across the Canadian Prairies, with climate scenarios from a Canadian RCM (CanRCM4) driven by a Canadian earth system model (CanESM2) with the forcing scenarios RCP4.5 and RCP8.5 at 25 and 50 km resolutions. Bias correction and a stochastic weather generator referred to as AAFC-WG were used as statistical processing tools to develop future climate scenarios as input to the crop model. The results showed that when changes were averaged across the locations, whether 25- or 50-km resolution CanRCM4 data were used, the projected yield changes were fairly consistent with those based on its driving GCM–CanESM2, especially if AAFC-WG was used to develop future climate scenarios. The spatial patterns of the projected yield changes were also similar for the two resolutions of CanRCM4, although the magnitude of the projected changes had a relatively large range among the scenarios at some locations. These results indicated that using future climate scenarios based on climate change simulations by GCMs might be sufficient for projecting regional crop yield changes on the Canadian Prairies.