Sensitivity of the pseudo-global warming method under flood conditions: A case study from the Northeastern U.S.

Abstract. Intensified extreme precipitation and resulting flooding are among the most impactful consequences of climate change, especially over the northeastern US (NEUS). To project and understand the impacts of climate change (or related climate perturbations) on extreme weather events as they may occur in the future, the Pseudo-Global Warming (PGW) method has been employed with great success. However, it has never been ascertained to what degree the conclusions from PGW studies are sensitive to the design of the PGW experiment. Consequently, three key questions related to the application of the PGW method remain unanswered: At what spatial scale should climate perturbations be applied? Among the different meteorological variables available, which should be perturbed? And will PGW projections vary significantly under different experiment designs? To begin to address these questions, we examine the sensitivity and robustness of conclusions drawn from the PGW method over NEUS by conducting multiple PGW experiments. The results show that the projections of precipitation and other essential variables are consistent at the regional mean scale, with a relative difference of much less than 10\%; however, different experimental designs nonetheless cause significant displacements among storm events. Several previously assumed advantages of modifying temperature at the regional mean scale do not appear to hold, such as the preservation of geostrophic balance. Also, we find the regional mean perturbation produces a positive precipitation bias due to overestimated warming over the ocean. Overall, PGW experiments with perturbations from temperature or the combination of temperature and wind at the gridpoint scale are both recommended, depending on the research target. The first approach can isolate the spatially-dependent thermodynamic impact, and the latter incorporates both the thermodynamic and dynamic impacts.