Klimamodelle fallen im Praxistest glatt durch: Ist die atmosphärische Zirkulation überhaupt simulierbar?

Ein Großteil der internationalen Klimapolitik fußt auf den Prognosen von Klimamodellen. Die Akteure tun dabei so, als wenn diese höchst robust seien und daher eine gute Entscheidungsgrundlage darstellen. Was dabei kaum durch den Presseschutzwall dringt, ist die hektische Diskussion der Klimamodellierer hinter den Kulissen. Theodore Shepherd von der University of Reading fasste im September 2014 das ganze Ausmaß der Probleme in einem Artikel in Nature Geoscience zusammen. Die Modelle bekommen die atmosphärische Zirkulation einfach nicht hin. Und selbst für die Zukunft könnte dies wohl so bleiben, befürchtet Shepherd:

Atmospheric circulation as a source of uncertainty in climate change projections
The evidence for anthropogenic climate change continues to strengthen, and concerns about severe weather events are increasing. As a result, scientific interest is rapidly shifting from detection and attribution of global climate change to prediction of its impacts at the regional scale. However, nearly everything we have any confidence in when it comes to climate change is related to global patterns of surface temperature, which are primarily controlled by thermodynamics. In contrast, we have much less confidence in atmospheric circulation aspects of climate change, which are primarily controlled by dynamics and exert a strong control on regional climate. Model projections of circulation-related fields, including precipitation, show a wide range of possible outcomes, even on centennial timescales. Sources of uncertainty include low-frequency chaotic variability and the sensitivity to model error of the circulation response to climate forcing. As the circulation response to external forcing appears to project strongly onto existing patterns of variability, knowledge of errors in the dynamics of variability may provide some constraints on model projections. Nevertheless, higher scientific confidence in circulation-related aspects of climate change will be difficult to obtain. For effective decision-making, it is necessary to move to a more explicitly probabilistic, risk-based approach.

Auch bei der Darstellung der Sonneneinstrahlung gibt es große Probleme, wie Zhou et al. 2015 zu bedenken gaben:

On the incident solar radiation in CMIP5 models
Annual incident solar radiation at the top of atmosphere should be independent of longitudes. However, in many Coupled Model Intercomparison Project phase 5 (CMIP5) models, we find that the incident radiation exhibited zonal oscillations, with up to 30 W/m2 of spurious variations. This feature can affect the interpretation of regional climate and diurnal variation of CMIP5 results. This oscillation is also found in the Community Earth System Model. We show that this feature is caused by temporal sampling errors in the calculation of the solar zenith angle. The sampling error can cause zonal oscillations of surface clear-sky net shortwave radiation of about 3 W/m2 when an hourly radiation time step is used and 24 W/m2 when a 3 h radiation time step is used.

Derzeit finden sich die Autorenteams für den geplanten 6. IPCC Klimazustandsbericht zusammen. Sind die schwerwiegenden Probleme mit den Klimamodellen nun endlich ausgeräumt? Keine Spur. Die Stony Brook University schlug am 11. Oktober 2017 Alarm: Die Modelle laufen immer noch nicht rund! Die deutsche Presse schweigt betreten. Hier die Stony Brook-Pressemitteilung:

Study Reveals Need for Better Modeling of Weather Systems for Climate Prediction
Computer-generated models are essential for or scientists to predict the nature and magnitude of weather systems, including their changes and patterns. Using 19 climate models, a team of researchers led by Professor Minghua Zhang of the School of Marine and Atmospheric Sciences at Stony Brook University, discovered persistent dry and warm biases of simulated climate over the region of the Southern Great Plain in the central U.S. that was caused by poor modeling of atmospheric convective systems – the vertical transport of heat and moisture in the atmosphere. Their findings, to be published in Nature Communications, call for better calculations in global climate models.  

The climate models analyzed in the paper “Causes of model dry and warm bias over central U.S. and impact on climate projections,” included a precipitation deficit that is associated with widespread failure of the models in capturing actual strong rainfall events in summer over the region. By correcting for the biases, the authors found that future changes of precipitation over the US Southern Great Plain by the end of the 21st Century would be nearly neutral. This projection is unlike what has been predicted as a drying period by the majority of current climate models. The correction also reduces the projected warming of the region by 20 percent relative to projections of previous climate models.

“Current climate models are limited by available computing powers even when cutting-edge supercomputers are used,” said Professor Zhang. “As a result, some atmospheric circulations systems cannot be resolved by these models, and this clearly impacts the accuracy of climate change predictions as shown in our study.” Professor Zhang and colleagues believe climate models will become more accurate in the coming years with the use of exsascale supercomputing, now in development worldwide.

Bereits 2014 hatten Mauri und Kollegen enorme Diskrepanzen zwischen realer und simulierter Entwicklung der Niederschläge und Temperaturen in Europa vor 5000 Jahren beanstandet. Die Rückwärtsmodellierung, also die Kalibrierung funktioniert überhaupt nicht. Bei soviel Enttäuschung darf die Frage erlaubt sein, woher eigentlich die Zuversicht kommt, dass dieselben Modelle die Zukunft verlässlich vorherzusagen vermögen. Hier der Abstract:

The influence of atmospheric circulation on the mid-Holocene climate of Europe: a data–model comparison
The atmospheric circulation is a key area of uncertainty in climate model simulations of future climate change, especially in mid-latitude regions such as Europe where atmospheric dynamics have a significant role in climate variability. It has been proposed that the mid-Holocene was characterized in Europe by a stronger westerly circulation in winter comparable with a more positive AO/NAO, and a weaker westerly circulation in summer caused by anti-cyclonic blocking near Scandinavia. Model simulations indicate at best only a weakly positive AO/NAO, whilst changes in summer atmospheric circulation have not been widely investigated. Here we use a new pollen-based reconstruction of European mid-Holocene climate to investigate the role of atmospheric circulation in explaining the spatial pattern of seasonal temperature and precipitation anomalies. We find that the footprint of the anomalies is entirely consistent with those from modern analogue atmospheric circulation patterns associated with a strong westerly circulation in winter (positive AO/NAO) and a weak westerly circulation in summer associated with anti-cyclonic blocking (positive SCAND). We find little agreement between the reconstructed anomalies and those from 14 GCMs that performed mid-Holocene experiments as part of the PMIP3/CMIP5 project, which show a much greater sensitivity to top-of-the-atmosphere changes in solar insolation. Our findings are consistent with data–model comparisons on contemporary timescales that indicate that models underestimate the role of atmospheric circulation in recent climate change, whilst also highlighting the importance of atmospheric dynamics in explaining interglacial warming.