State of a challenge – Third annual review

We announced a public challenge at the EGU 2020 General Assembly against CMIP6 models predicting an increase of downward longwave radiation (DLR) in the range of 10 – 40 Wm^-2 during the 21^st century as a result of human greenhouse gas emissions. We based our challenge on observed facts, supported by long-known but rarely referred theoretical constraints. 22 years of CERES data show +0.11 Wm^-2/decade increase in DLR, equivalent to +0.36 Wm^-2 increase (+0.06 °K) until 2050 (in contrast to IPCC AR6, predicting +2 Wm^-2/decade).

Supporting our prediction, we repeat here the deduction of the constraint equations, and control them on the recently available data sets. — Our best tool the compute the transfer of radiation in the atmosphere is Schwarzschild’s (1914) equation; its early, two-stream form is given in Schwarzschild (1906, Eq. 11), appropriate for global-mean energy flow computations. The equation consists of three terms; the difference of the second and first terms gives the net radiation at the surface as constrained to half of the outgoing longwave radiation (OLR), independently of the optical depth.  In the literature it was observed early (Emden 1913) that there is a discontinuity at the surface in radiative equilibrium, balanced by the turbulent fluxes in radiative-convective equilibrium. The formula for this net radiation is given for example in the textbook of Goody (1964, Atmospheric radiation: theoretical basis); repeated by Houghton (1977, Eq. 2.13), graphically represented in Chamberlain (1979, Fig. 1.4); and verified by the data (without explicitly describing the equation) of Hartmann (1994, pp. 61-63) within 0.3 Wm^-2. The equation is verified by the CERES EBAF Ed2.8 (16 years of clear-sky global mean data) within 0.6 Wm^-2. We use the second term of Schwarzschild (1906, Eq.11) with a particular optical depth of τ = 2 to compute the total energy absorption (and emission) at the surface, verified by the same satellite data product within the same difference (0.6 Wm^-2) in the clear-sky annual global mean. — We created the all-sky versions of these two equations by introducing longwave cloud radiative effect (LWCRE), and justified the four individual equations on the most recent 22 years of CERES EBAF Edition 4.1 global mean data within ±3 Wm^-2; while the mean bias of the four equations together is 0.0007 Wm^-2. These equations form the boundary conditions of every valid climate prediction.

Reference:
Zagoni, M.: Challenging CMIP6 model predictions, EGU General Assembly 2020, Online, 4–8 May 2020, EGU2020-1, https://doi.org/10.5194/egusphere-egu2020-1