Attribution of historical climate changes is particularly challenging for periods before the mid-20th century due to sparse and low-quality observational data, raising questions about which temporal scales best capture the climate system’s response to external forcings and can robustly constrain future projections. To address these issues, this study performs a multi-scale attribution analysis of temperature extremes over the period 1901–2020 and examines the robustness of attribution-based scaling factors across different time scales. Despite uncertainties in the early data, the newly developed homogenized observations show pronounced warming in both cold and hot extremes, along with a lengthening of the growing season during 1901–2020. These trends intensified markedly after the 1950s, with the magnitude of changes approximately doubling for some extreme indices. Coupled Model Intercomparison Project Phase 6 (CMIP6) models successfully reproduce the overall warming trends in observations, although they underestimate the magnitude of changes, particularly in the pre-1950 period. Using optimal fingerprinting, more than 70% of the observed changes are attributed to greenhouse gas forcing, with aerosols offsetting less than 35% of the greenhouse gas-induced warming. Attribution analysis conducted within a large-ensemble model framework across multiple time scales shows that the ranges of best estimates and confidence intervals (CIs) for scaling factors decrease as the time period lengthens. The century-scale attribution yields the narrowest CIs and most robust best estimates, indicating the most robust detection results. Despite the robustness of century-scale results, scaling factors from 1951–2020 are selected to constrain projections due to more reliable observational constraints. Constrained end-of-century (2081–2100) projections show amplified increases of 20.3%–33.1% for most extremes compared to raw projections, highlighting the critical impact of attribution period selection and providing a transferable framework for regional climate risk assessment.

Read original article

Read original article