4 ms·

> Which still overpredict... Overpredict? According to what? According to NASA[1] we're already at a 0.9 degree anomaly for 2017. The IPCC prediction[2] is 1

> Which still overpredict... Overpredict? According to what? According to NASA[1] we're already at a 0.9 degree anomaly for 2017. The IPCC prediction[2] is 1 degree anomaly in 2020 for the worst case prediction in the worst case scenario. The NASA anomaly is relative to 1951-1980 average while IPCC is 1986-2005 average. So it's not a perfect comparison. But if you want to look at IPCC's best case scenarios, the prediction is about 0.1 degree, which certainly refutes your claim that they overpredict. [1]: https://climate.nasa.gov/vital-signs/global-temperature/ https://climate.nasa.gov/vital-signs/global-temperature/ [2]: https://www.ipcc.ch/pdf/assessment-report/ar5/syr/AR5_SYR_FINAL_SPM.pdf https://www.ipcc.ch/pdf/assessment-report/ar5/syr/AR5_SYR_FI... See page 11

pdonis 8y ago
> Overpredict? According to what? See, for example, here: https://judithcurry.com/2015/12/17/climate-models-versus-climate-reality/ https://judithcurry.com/2015/12/17/climate-models-versus-cli... Note that this actually understates the overprediction, because the models whose average is shown included three sets of models that made three different assumptions about how CO2 would rise. But CO2's actual rise over the period where we're comparing models with data was about the same as the assumption of the "worst case" (or "business as usual") scenario, so that scenario is the only one that's relevant to the comparison between models and data. And when you compare just that sheaf of models with the data, the divergence between model and data is even larger. For the reasons why the upper troposphere temperature comparison is more relevant than the surface temperature comparison (which is what is shown in the IPCC AR5), see the article. > According to NASA[1] we're already at a 0.9 degree anomaly for 2017. The IPCC prediction[2] is 1 degree anomaly in 2020 for the worst case prediction in the worst case scenario. But, as you note, the two anomalies are relevant to two different baselines. That means the comparison is invalid, not "not perfect". > if you want to look at IPCC's best case scenarios, the prediction is about 0.1 degree That's for future temperature change under the future best case scenario. It has nothing to do with past temperature change under the actual scenario that actually happened. > See page 11 Those graphs don't show models vs. data. They show models only (and with the actual time period for which we have data compressed on the far left of the graphs so you can't extract any useful comparison with data from them). The Summary for Policymakers, which is what you linked to, does not show any direct comparison between models and data for a period for which we have data. Previous reports did, but in this one they decided not to--it's only buried in the WG1 report, in Chapter 9, p. 768: http://www.ipcc.ch/pdf/assessment-report/ar5/wg1/WG1AR5_Chapter09_FINAL.pdf http://www.ipcc.ch/pdf/assessment-report/ar5/wg1/WG1AR5_Chap... The divergence between the CMIP5 mean (red) and the actual temperatures (black) in the upper graph in the 2000s is visible; the EMIC mean in the lower graph only appears to go up to 2000 so it doesn't provide a useful comparison. Also, both of these graphs show all of the models, not just the ones whose CO2 rise assumptions matched reality over the data period, so they understate the actual overprediction for the reason I gave above.

pdonis 8y ago

> Overpredict? According to what? See, for example, here: https://judithcurry.com/2015/12/17/climate-models-versus-climate-reality/ https://judithcurry.com/2015/12/17/climate-models-versus-cli... Note that this actually understates the overprediction, because the models whose average is shown included three sets of models that made three different assumptions about how CO2 would rise. But CO2's actual rise over the period where we're comparing models with data was about the same as the assumption of the "worst case" (or "business as usual") scenario, so that scenario is the only one that's relevant to the comparison between models and data. And when you compare just that sheaf of models with the data, the divergence between model and data is even larger. For the reasons why the upper troposphere temperature comparison is more relevant than the surface temperature comparison (which is what is shown in the IPCC AR5), see the article. > According to NASA[1] we're already at a 0.9 degree anomaly for 2017. The IPCC prediction[2] is 1 degree anomaly in 2020 for the worst case prediction in the worst case scenario. But, as you note, the two anomalies are relevant to two different baselines. That means the comparison is invalid, not "not perfect". > if you want to look at IPCC's best case scenarios, the prediction is about 0.1 degree That's for future temperature change under the future best case scenario. It has nothing to do with past temperature change under the actual scenario that actually happened. > See page 11 Those graphs don't show models vs. data. They show models only (and with the actual time period for which we have data compressed on the far left of the graphs so you can't extract any useful comparison with data from them). The Summary for Policymakers, which is what you linked to, does not show any direct comparison between models and data for a period for which we have data. Previous reports did, but in this one they decided not to--it's only buried in the WG1 report, in Chapter 9, p. 768: http://www.ipcc.ch/pdf/assessment-report/ar5/wg1/WG1AR5_Chapter09_FINAL.pdf http://www.ipcc.ch/pdf/assessment-report/ar5/wg1/WG1AR5_Chap... The divergence between the CMIP5 mean (red) and the actual temperatures (black) in the upper graph in the 2000s is visible; the EMIC mean in the lower graph only appears to go up to 2000 so it doesn't provide a useful comparison. Also, both of these graphs show all of the models, not just the ones whose CO2 rise assumptions matched reality over the data period, so they understate the actual overprediction for the reason I gave above.