Mohamed Ali, Ben Alaya, Francis W. Zwiers, Xuebin Zhang (2022). An extreme value based likelihood ratio test to evaluate BCCAQv2’s capability for downscaling and projecting future unprecedented precipitation extremes. Proceedings of the GWF Annual Open Science Meeting, May 16-18, 2022.

The spotlight on global warming is progressively turning away from ordinary events toward understanding and assessing the consequences on extreme events. In Canada, the principal source of downscaled future precipitation is statistical downscaling using the Bias Correction Constructed Analogues with Quantile mapping reordering version 2 (BCCAQv2) approach. Despite the widespread use of this information, our proper understanding of its trustworthiness remains relatively restricted. Indeed, the downscaling process's reliability is contingent upon the BCCAQv2 transfer function being stable in a future warmer climate. Observational data, on the other hand, cannot be used to test this premise since we lack "future observations”. Operating in a "perfect model" world is, nonetheless, a comprehensive way to investigate this hypothesis. We therefore assess whether the BCCAQv2 transfer function is capable of downscaling precipitation extremes from a conventional Canadian Regional Climate Model CanRCM4 to a convection-permitting (CP) 4-kilometer resolution across southern British Columbia, Canada. After calibrating the BCCAQv2 transfer function using historical CanRCM4 precipitation simulations and a retrospective simulation of the Weather Research Forecasting (WRF) model (CTL, 2000-2014) as a target, a pseudo-global warming (PGW) WRF simulation of future climate is used to verify whether the BCCAQv2 transfer function is affected by climate change. Because of the inherent rarity of extreme events, any statistical estimate of their impacts will be fraught with doubt. Here we propose a likelihood ratio test underpinned by the statistical theory of extreme values to examine BCCAQv2's potential to generate unprecedented precipitation extremes under both current and future climatic circumstances. Main results suggests that we should think a little more deeply about the impact of using a transfer function that is strongly affected by warming signals and how to reduce its related biases.

AOSM2022 Climate-Related Precipitation Extremes First Author: Mohamed Ali Ben Alaya, Pacific Climate Impacts Consortium, University of Victoria Additional Authors: Francis W. Zwiers, Pacific Climate Impacts Consortium, University of Victoria / Xuebin Zhang, Climate Research Division, Environment and climate change Canada