Japan's boffins: Global warming isn't man-made
Climate science is 'ancient astrology', claims report
Predicting the Future with Numerical Simulation
Kanya Kusano, Japan Agency for Marine-Earth Science & Technology (JAMSTEC)
Numerical simulation by forecast models are generally classified as theoretical models and empirical models. The former follows universal laws and carries out predictive calculations, the latter makes models that are thought to be realistic from data of phenomenon. These two methods cannot be strictly differentiated, generally experiential methods gradually become theoretical methods, finally becoming the generally accepted dogma.
Celestial mechanics originated in astrological prediction of solar and lunar eclipses, calendars were experiential predictions; mechanistic theory evolved when we reached an era of accurate computation. Consequently, the predictability of celestial mechanics became extremely high and practical estimates gave way to proof. Similarly, modern Global Climate Models still largely dependent on empirical models. Fundamental principles, therefore must resolve very complex physical/chemical/biological processes and phenomenon. That is why many artificial optimization operations (parameterization tuning) are needed, or we will not be able to reproduce the phenomenon. Because of this, besides mathematical accuracy, the people who construct models' choice of processes and optimum operating guidelines will have large scale effects on the calculated results.
1. Scientific Understanding and Uncertainty
When constructing models, if our scientific understanding is poor, we are not able to capture the model. But we should pay attention to the importance of the naturally occurring processes when our scientific understanding is not yet clearly decided.
In the IPCC's 4th Evaluation Report, a few potentially major processes were discussed; but [since] scientific understanding was too low to decide, the evaluation of these was omitted. In order to scientifically understand the uncertainty of accurate estimates according to the potential importance of these processes, "the cause of lack of scientific understanding and uncertainty" must be assessed.
Finally, uncertainty estimates should be included. For example, the effect of variances in cosmic ray activity on clouds, caused by sunspot activity, solar flares accompanied by energetic protons striking the upper atmosphere and generating NOx and ozone effects [*], etc., are not sufficiently understood and incorporated into the models.
Also, there are great uncertainties in reproducing historical TSI (Total Solar Irradiance), TSI fluctuation and spectral change related climate sensitivity estimates are inadequate.
Sponsored: Hyper-scale data management