Gavin A. Schmidt, Drew T. Shindell & Kostas Tsigaridis have a new article in Nature Geoscience (open access). What they've done is estimate the impact of actual measures of solar, volcanoes and ENSO on a CMIP5 ensemble. They found that this reduced the recent difference between models and observations a whole lot.
What they found in particular was the the models most likely overestimated the cooling from the Pinatubo eruption in the 1990s, making the models too cool and, when observed solar radiation, volcanic eruptions and ENSO were factored, in the models are pretty close to observations.
Here is the figure from the paper. Click to enlarge it.
 |
| Figure 1: Updated external influences on climate and their impact on the CMIP5 model runs. a, The latest reconstructions of optical depth for volcanic aerosols9, 10 from the Mount Pinatubo eruption in 1991 suggest that the cooling effect of the eruption (1991–1993) was overestimated in the CMIP5 runs, making the simulated temperatures too cool. From about 1998 onwards, however, the cooling effects of solar activity (red), human-made tropospheric aerosols (green) and volcanic eruptions (pink) were all underestimated. WMGHG, well-mixed greenhouse gases. b, Global mean surface temperature anomalies, with respect to 1980–1999, in the CMIP5 ensemble (mean: solid blue line; pale blue shading: 5–95% spread of simulations) on average exceeded two independent reconstructions from observations (GISTEMP Land–Ocean Temperature Index (LOTI)6, solid red; HadCRUT4 with spatial infilling7, dashed red) from about 1998. Adjusting for the phase of ENSO by regressing the observed temperature against the ENSO index11 adds interannual variability to the CMIP5 ensemble mean (dashed blue), and adjusting for updated external influences as in a further reduces the discrepancy between model and data from 1998 (black). The adjusted ensemble spread (dashed grey) clearly shows the decadal impact of the updated drivers. As an aside, we note that although it is convenient to use the CMIP5 ensemble to assess expected spreads in possible trends, the ensemble is not a true probabilistic sample. |
It's going to get hotter
The authors conclude the following, which won't come as news (or welcome news) to HotWhopper readers (my bold):
We conclude that use of the latest information on external influences on the climate system and adjusting for internal variability associated with ENSO can almost completely reconcile the trends in global mean surface temperature in CMIP5 models and observations. Nevertheless, attributing climate trends over relatively short periods, such as 10 to 15 years, will always be problematic, and it is inherently unsatisfying to find model–data agreement only with the benefit of hindsight. We see no indication, however, that transient climate response is systematically overestimated in the CMIP5 climate models as has been speculated, or that decadal variability across the ensemble of models is systematically underestimated, although at least some individual models probably fall short in this respect.
Most importantly, our analysis implies that significant warming trends are likely to resume, because the dominant long-term warming effect of well-mixed greenhouse gases continues to rise. Asian pollution levels are likely to stabilize and perhaps decrease, although lower solar activity may persist and volcanic eruptions are unpredictable. ENSO will eventually move back into a positive phase and the simultaneous coincidence of multiple cooling effects will cease. Further warming is very likely to be the result.
Anthony Watts hasn't picked up on this paper yet, but I expect he will sooner or later.
For the record, this is a link to the March 2014 special issue of Nature Geoscience "Recent slowdown in global warming".
Gavin A. Schmidt, Drew T. Shindell & Kostas Tsigaridis, Reconciling warming trends, Nature Geoscience 7, 158–160 (2014) doi:10.1038/ngeo2105 Published online 27 February 2014
