The IPCC AR6 assessment of the impacts and risks associated with projected climate changes for the 21st century is both alarming and ambiguous. According to computer projections, global surface temperature may warm from 1.3 degrees C to 8.0 degrees C by 2100, depending on the global climate model (GCM) and the shared socioeconomic pathway (SSP) scenario used for the simulations. Actual climate-change hazards are estimated to be high and very high if the global surface temperature rises, respectively, more than 2.0 degrees C and 3.0 degrees C above pre-industrial levels. Recent studies, however, showed that a substantial number of CMIP6 GCMs run "too hot" because they appear to be too sensitive to radiative forcing, and that the high/extreme emission scenarios SSP3-7.0 and SSP5-8.5 are to be rejected because judged to be unlikely and highly unlikely, respectively. Yet, the IPCC AR6 mostly focused on such alarmistic scenarios for risk assessments. This paper examines the impacts and risks of "realistic" climate change projections for the 21st century generated by assessing the theoretical models and integrating them with the existing empirical knowledge on global warming and the various natural cycles of climate change that have been recorded by a variety of scientists and historians. This is achieved by combining the SSP2-4.5 scenario (which is the most likely SSP according to the current policies reported by the International Energy Agency) and empirically optimized climate modeling. According to recent research, the GCM macro-ensemble that best hindcast the global surface warming observed from 1980 to 1990 to 2012-2022 should be made up of models that are characterized by a low equilibrium climate sensitivity (ECS) (1.5 degrees C < ECS <= 3.0 degrees C), in contrast to the IPCC AR6 likely and very likely ECS ranges at 2.5-4.0 degrees C and 2.0-5.0 degrees C, respectively. I show that the low-ECS macro-GCM with the SSP2-4.5 scenario projects a global surface temperature warming of 1.68-3.09 degrees C by 2080-2100 instead of 1.98-3.82 degrees C obtained with the GCMs with ECS in the 2.5-4.0 degrees C range. However, if the global surface temperature records are affected by significant non-climatic warm biases - as suggested by satellite-based lower troposphere temperature records and current studies on urban heat island effects - the same climate simulations should be scaled down by about 30%, resulting in a warming of about 1.18-2.16 degrees C by 2080-2100. Furthermore, similar moderate warming estimates (1.15-2.52 degrees C) are also projected by alternative empirically derived models that aim to recreate the decadal-to-millennial natural climatic oscillations, which the GCMs do not reproduce. The proposed methodologies aim to simulate hypothetical models supposed to optimally hindcast the actual available data. The obtained climate projections show that the expected global surface warming for the 21st-century will likely be mild, that is, no more than 2.5-3.0 degrees C and, on average, likely below the 2.0 degrees C threshold. This should allow for the mitigation and management of the most dangerous climate-change related hazards through appropriate low-cost adaptation policies. In conclusion, enforcing expensive decarbonization and net-zero emission scenarios, such as SSP1-2.6, is not required because the Paris Agreement temperature target of keeping global warming < 2 degrees C throughout the 21st century should be compatible also with moderate and pragmatic shared socioeconomic pathways such as the SSP2-4.5. (c) 2023 China University of Geosciences (Beijing) and Peking University. Published by Elsevier B.V. on behalf of China University of Geosciences (Beijing). This is an open access article under the CC BY-NC-ND license .
