by Judith Curry
This lack of precise knowledge of surface energy fluxes profoundly affects our ability to understand how Earth’s climate responds to increasing concentrations of greenhouse gases. – Graeme Stephens et al.
An update on Earth’s energy balance in light of the latest global observations
Graeme L. Stephens, Juilin Li, Martin Wild, Carol Anne Clayson, Norman Loeb, Seiji Kato, Tristan L’Ecuyer, Paul W. Stackhouse Jr, Matthew Lebsock and Timothy Andrews
Abstract. Climate change is governed by changes to the global energy balance. At the top of the atmosphere, this balance is monitored globally by satellite sensors that provide measurements of energy flowing to and from Earth. By contrast, observations at the surface are limited mostly to land areas. As a result, the global balance of energy fluxes within the atmosphere or at Earth’s surface cannot be derived directly from measured fluxes, and is therefore uncertain. This lack of precise knowledge of surface energy fluxes profoundly affects our ability to understand how Earth’s climate responds to increasing concentrations of greenhouse gases. In light of compilations of up-to-date surface and satellite data, the surface energy balance needs to be revised. Specifically, the longwave radiation received at the surface is estimated to be significantly larger, by between 10 and 17 Wm−2, than earlier model-based estimates. Moreover, the latest satellite observations of global precipitation indicate that more precipitation is generated than previously thought. This additional precipitation is sustained by more energy leaving the surface by evaporation — that is, in the form of latent heat flux — and thereby offsets much of the increase in longwave flux to the surface.
Citation: Nature Geoscience 5, 691–696 (2012) doi:10.1038/ngeo1580 [link]
The punchline of the paper is this Figure, which is essentially a re-do of the Kiehl and Trenberth figure:
The models are hunting for imbalances and build-ups in planetary energy. But according to the observations, the longwave (infra-red) energy coming onto the earth’s surface, the infamous back radiation, is 10 – 17 W/m2 higher than in the famous Trenberth diagram from 1997. So the models are trying to explain tiny residual imbalances, but the uncertainties and unknowns are larger than the target. The argument that “only the forcing from CO2 can fill the gap in the models” is not just argument from ignorance rhetorically, but factually too.
Another major implications is that water is churning up and falling out of the sky faster than the experts thought. The Earth’s evaporative cooler is lifting more water, taking more heat, and dumping that heat in the atmosphere. At the top of the atmosphere heat is radiating off the planet to offset the radiation coming in. On the water planet, it really is all about water.
Abstract. Four different types of estimates of the surface downwelling longwave radiative flux (DLR) are reviewed. One group of estimates synthesizes global cloud, aerosol, and other information in a radiation model that is used to calculate fluxes. Because these synthesis fluxes have been assessed against observations, the global-mean values of these fluxes are deemed to be the most credible of the four different categories reviewed. The global, annual mean DLR lies between approximately 344 and 350 W m−2 with an error of approximately ±10 W m−2 that arises mostly from the uncertainty in atmospheric state that governs the estimation of the clear-sky emission. The authors conclude that the DLR derived from global climate models are biased low by approximately 10 W m−2 and even larger differences are found with respect to reanalysis climate data. The DLR inferred from a surface energy balance closure is also substantially smaller that the range found from synthesis products suggesting that current depictions of surface energy balance also require revision. The effect of clouds on the DLR, largely facilitated by the new cloud base information from the CloudSat radar, is estimated to lie in the range from 24 to 34 W m−2 for the global cloud radiative effect (all-sky minus clear-sky DLR). This effect is strongly modulated by the underlying water vapor that gives rise to a maximum sensitivity of the DLR to cloud occurring in the colder drier regions of the planet. The bottom of atmosphere (BOA) cloud effect directly contrast the effect of clouds on the top of atmosphere (TOA) fluxes that is maximum in regions of deepest and coldest clouds in the moist tropics.
First, kudos to Stephens et al. for producing a very important analysis, which has clearly required a great deal of effort and was made possible by the new satellite data products. Providing justified uncertainty estimates on these numbers is an important step forward.
So what are we to make of this in terms of climate models and attribution of 20th century warming? Stephens et al. compare their analysis with the CMIP5 model simulations:
Models are commonly tuned to the TOA, so direct comparison of TOA fluxes provides little insight into model performance. As the surface solar flux is also correlated to the TOA reflected solar flux, that flux is also not entirely free of ‘tuning’ effects, so a direct comparison with estimated surface solar flux also has to be interpreted cautiously. The remaining surface fluxes, however, are completely uncoupled from the TOA fluxes and comparison with observations reveals important insights about model energy balances. The model fluxes given in Fig. 1 are expressed as a multi-model average and a range indicated by maxima and minima fluxes of the model ensemble. The inter-model global mean fluxes lie within the uncertainty of the observed values, and the global mean downward longwave surface fluxes taken from climate models generally lie at the low end of the uncertainty range of the estimated fluxes as noted in other studies. It is also notable that the model latent heat fluxes are closer to the new revised flux. Although model and observations broadly agree in the global mean, important regional biases exist in the modelled energy budgets that are not conveyed in global mean statistics.
Note, concluding that climate models are incorrect because of this new analysis of the global heat budget is NOT justified. The Kiehl-Trenberth diagram is not used in climate models in any way, and mainly has been used as a conceptual aid. The CMIP5 models actually agree better with the Stephens et al. analysis than with earlier analyses. That said, the Stephens et al. analysis highlights the uncertainties in our ability to observe and simulate the global mean surface energy balance.
The really interesting issue is the variability: regional and temporal. I hope that this data set will be used in future studies that addresses these issues.
Moderation note: This is a technical thread, keep your comments relevant and on topic.