by Judith Curry, February 13, 2020 in WUWT
A range of scenarios for global mean surface temperature change between 2020 and 2050, derived using a semi-empirical approach. All three modes of natural climate variability – volcanoes, solar and internal variability – are expected to act in the direction of cooling during this period.
In the midst of all the angst about 1.5oC or 2.0oC warming or more, as defined relative to some mythical time when climate was alleged to be ‘stable’ and (relatively) uninfluenced by humans, we lose sight of the fact that we have a better baseline period – now. One advantage of using ‘now’ as a baseline for future climate change is that we have good observations to describe the climate of ‘now’.
While most of the focus of climate projections is on 2100, the period circa 2020-2050 is of particular importance for several reasons:
It is the period for meeting UNFCCC targets for emissions reductions
Many financial and infrastructure decisions will be made on this time scale
The actual evolution of the climate over this period will influence 1) and 2) above; ‘surprises’ could have adverse impacts on decisions related to 1) and 2).
Figure 1: CO2-induced warming as a function of cumulative emissions and TCRE. Millar et al
Global climate/earth system models have little skill on decadal time scales. To address this issue, CMIP5 and CMIP6 are conducting initialized, decadal scale simulations out to 35 years. While I haven’t seen any CMIP6 decadal results yet, I do follow this literature. Punchline is that there is some skill in simulating the Atlantic Multidecadal Oscillation (AMO) out to 8-10 years, but otherwise not much overall skill.
I have previously criticized the interpretation CMIP5 simulations as actual climate change scenarios – instead, these simulations show the sensitivity of climate to different emissions scenarios. They neglect scenarios of future solar variability, volcanic eruptions, and the correct phasing and amplitude of multidecadal variability associated with ocean circulations. The argument for dismissing these factors is that they are smaller than emissions forcing. Well, cumulatively and on decadal to multi-decadal timescales, this is not necessarily true.
And in the CMIP6 era, we now have sufficient information and understanding so that we can generate plausible scenarios of volcanic and solar forcing for the 21st century, as well as for the AMO.
I have developed a semi-empirical approach to formulating 21st century climate change scenarios that rely only indirectly on climate models. Multiple scenarios are generated for each driver of the forecast (natural and anthropogenic), with an emphasis on plausible scenarios (rather than extreme scenarios that cannot completely be ruled out).
Note: in what follows, many references are cited. I don’t have time now to pull together a full bibliography, but I have provided hyperlinks to the key references.