Climate model

from Wikipedia, the free encyclopedia
Terra X : This is how climate models work

A climate model is a computer model for calculating and projecting the climate for a certain period of time. The model is usually based on a meteorological model , as it is also used for numerical weather forecast . However, this model is extended for climate modeling in order to correctly map all conserved quantities. Usually an ocean model , a snow and ice model for the cryosphere and a vegetation model for the biosphere are coupled.

Mathematically, this creates a coupled system of non-linear, partial and ordinary differential equations as well as some algebraic equations. The numerical calculation of this system of equations requires a great deal of computing power, as provided by supercomputers such as the Earth Simulator .

A distinction is made between global climate models (so-called GCMs, general circulation models) and regional climate models. The main difference is, on the one hand, that a global climate model includes the entire troposphere , while a regional model usually depicts the same model physics, but only applies this to a certain geographical section of the earth.

General

Climate models represent the most complex and computationally expensive computer models that have been developed so far. The "extrapolations" of the climate models are naturally more uncertain than those of the weather models, since much longer periods of time have to be taken into account and a large number of additional parameters must be taken into account. For this reason, these individual models are also called climate scenarios and not climate predictions. A weather forecast is based on data, which makes it possible to predict the development of the chaotic dynamics within the earth's atmosphere with a high probability within a period of currently up to a week. The uncertainty of the extrapolation increases exponentially with the extrapolated period and is itself dependent, among other things, on the weather situation. Even with weather models, the experience and assessment of the user in the form of a control instance between the pure computer model and the ultimate forecast play a decisive role, but the character of a climate model is fundamentally different.

Example of modeled and measured global average temperatures between 1900 and 2000.

Climate models are used to find possible trends in the development of the climate and to weight individual climate factors. They are based on a large number of assumptions and methods, for example on the development of future greenhouse gas emissions and feedback mechanisms. In addition, in contrast to weather models, climate models are not based on fixed dynamics and are therefore not restricted by their chaotic character and the limitation of computing power. Some feedbacks, especially in connection with the tilting elements in the earth system , have not yet been adequately researched and can not be adequately reconstructed, including the history of the climate .

The modeling, including suitable parameter values, is the subject of ongoing scientific work.

Role of climate models in simulating the course and consequences of global warming

To check whether the parameters with which climate models are calculated are correct, they are tested to see whether they can correctly simulate the current climate as well as the climate during the ice ages. In the context of such simulations, over 1000 models are calculated, with input parameters being varied within their assumed error range. Models that do not correctly reflect the temperature profile in the period under consideration (> 90%) are rejected.

Although knowledge of paleoclimatology has expanded dramatically since the 1980s, the data on climate history are still incomplete; Due to the speed as well as the extent of the global warming expected in the future , one is likely to enter "new territory" with partly unforeseen consequences .

The prerequisites necessary for the modeling are therefore only partially known and usually have to be determined more or less arbitrarily, whereby a set of these definitions and the modeling based on them are referred to as a climate scenario . The difference between a climate forecast and a climate scenario is that a large number of different scenarios are modeled for the former, on the one hand with other models and on the other with other anticipatory assumptions. A climate prognosis is based on the evaluation of various modeling experiments and is very difficult to make and with enormous effort, due to the difficult comparability between them. Since the individual scenarios, which are also reflected in the structure of the Intergovernmental Panel on Climate Change , show different final results, a climate forecast based on them can only reveal a range of possibilities. In the case of global warming , this range corresponds to a possible warming of the average global and ground-level air temperature of 1.1 to 6.4 ° C by the year 2100 (IPCC 2007). However, similar ranges of fluctuation can be found in almost all projections derived from climate models.

Global climate models - GCM (General Circulation Model)

A global climate model describes the most important climate-relevant physical processes in the earth's atmosphere, the oceans and on the earth's surface. The processes are shown in a very simplified way. Above all, the processes in the biosphere are currently still specified as sizes and parameters. However, these variables are system variables and should be able to adapt to global change during the simulation in order to be able to provide realistic projections for the future. Such feedback processes from coupled systems are currently the great challenge in modeling. The models are so extensive that they can only be operated in a very coarse resolution (several hundred kilometers grid width). The first GCM was created in 1967 by Syukuro Manabe and Richard Wetherald.

Examples of global climate models are:

  • HadCM3 (Hadley coupled model, version 3): This climate model was used, along with several others, for the third (TAR) and fourth (AR4) assessment reports of the IPCC
  • HadGEM1 (Hadley global environment model 1): Further development of the HadCM3 climate model. The representation of the influence of clouds and sea ice has been improved; The mapping of the following parameters has also been improved: water balance, atmospheric chemistry and the effects of aerosols. However, the representation of the influences of El Niño, monsoons, and Pacific surface temperatures have deteriorated and are the subject of ongoing research with progress being made.

Regional climate models

Regional climate models only consider a section of the atmosphere and therefore require suitable boundary conditions at the edges of the simulation area. These boundary conditions come from simulations of the global climate models. It is therefore said that a regional climate model is driven by a global climate model. This is known as “nesting” or “dynamic downscaling” and describes the embedding of a regional model with a high spatial resolution in a global climate model with a low spatial resolution. The distances between the grid points in a global climate model are usually quite large and are between 150 and 500 km. Regional models, on the other hand, have a very fine resolution. The grid points are sometimes only 1 km apart. Due to the increase in the computing capacity of modern supercomputers , the spatial resolution of the models can be continuously improved.

Examples of regional models are:

Climate modeling in Germany

In Germany, climate models are used for very different research questions at a large number of universities and research institutes. One of the central locations is the Max Planck Institute for Meteorology in Hamburg. Among other things, the global atmosphere models ECHAM -4 and ECHAM-5 and the ocean model MPI-OM were developed there. ECHAM and MPI-OM are used both as stand-alone components and coupled with one another, depending on the scientific question. The neighboring German Climate Computing Center (DKRZ) is closely linked to the MPI for Meteorology . There are parallel vector computers available, as are required for the operation of the models. The DKRZ is also available to other research institutions to operate these models, among other things.

The regional climate modeling is carried out in the large research institutes with various regional models. These research centers include the Karlsruhe Research Center , the GKSS Research Center in Geesthacht, the Potsdam Institute for Climate Impact Research (PIK) and several universities. The climate models are partly developed together with the German Weather Service. In addition to the continuous further development of numerical models, ensemble simulations and their probabilistic interpretation are gaining in importance.

The Federal Environment Agency and the Federal Institute for Hydrology (BfG) have had regional climate projections created for Germany up to the year 2100. The raw data of these model runs are available to the public free of charge.

Many current research projects deal with the question of the interactions between different subsystems of the climate system. Attempts are therefore made to integrate other subsystems into the climate models in addition to the atmosphere and ocean, for example biosphere or cryosphere . In this context, we speak of earth system models. (See also: Earth Systems Science .)

Research institutions that deal with the interactions between the atmosphere and the biosphere are for example the Max Planck Institute for Biogeochemistry (Jena), the Max Planck Institute for Meteorology (Hamburg) or the Potsdam Institute for Climate Impact Research . The Center for Marine Environmental Sciences (Bremen) researches the interactions between the biosphere and the physiosphere in the ocean.

At the Danmarks Meteorologiske Institut , the climate model HIRHAM5 (version 5 of the HIRHAM model) was developed in cooperation with scientists from the Alfred Wegener Institute , which is based on version 7.0 of the HIRLAM model and version 5.2.02 of the ECHAM model. Compared to the predecessor model HIRHAM4 based on ECHAM Version 4 and HIRLAM Version 2, a semi-Lagrange scheme is used in HIRHAM5 when modeling the dynamics, which allows larger time steps and at the same time a higher geographical resolution compared to the Euler method used before.

Limits of the climate models

When interpreting the results of the current climate model calculations in the future, it must be taken into account that these are not predictions about a secure future course of local or global climates, but rather scenarios which select possible courses based on assumptions about future developments, such as Example emissions and land use.

The limits of the models lie in the mathematical models used themselves and in the limited number of influencing factors considered. More powerful computers enable the development of more complex models with higher spatial resolutions and an increasing number of influencing factors on the climate. If the physical principles are only poorly understood, currently the case with the dynamics of ice sheets or the role of aerosols and clouds, climate models can only deliver comparatively uncertain results. For warm phases of the past 3.5 million years, model simulations show a global mean temperature that is up to 50% lower than palaeoclimatological data. This suggests that the climate models significantly underestimate the long-term temperature and sea level rise caused by the current warming.

In the ice cores of the Arctic, recurring abrupt climatic changes of a considerable extent are documented. These can only inadequately be reproduced with today's computer models. Richard B. Alley suspects that a number of feedback loops and side effects are not yet taken into account in the modeling.

An example of a failure of climate models is the unexpectedly high decline in Arctic sea ice cover, as it occurred in the summer of 2007. The sea ice shrinkage was the result of changed pressure and circulation patterns that have replaced the previous regime for several years. The possibility of such a development for the next few years was not shown in any of the climate models in the IPCC's climate report published in the same year.

Various blogs in the climate denial scene regard the pause in global warming that was allegedly registered between 1998 and around 2013 (actually only a stagnation in surface temperature after a very hot initial year 1998) as a sign of the failure of climate models to correctly predict the course of global warming . In science, however, reference is made to the factors cooling the global climate that were active in this period and that were not part of the models. In addition, there was a strong El Nino at the beginning of this period , which naturally results in global heat anomalies, which is why evaluations starting this year show a reduced trend. Accordingly, the reduced temperature rise disappears if the very warm years from 2014 to 2016 are included in the trend analysis. A study published in 2015 in the journal Nature came to a similar conclusion , which compared temperature increases determined in climate models with actually measured temperature increases between 1900 and 2012. Accordingly, there are no indications that climate models systematically overestimate the effects of greenhouse gases; rather, the deviation in the period from 1998 to 2012 is largely due to random statistical fluctuations and a small contribution from volcanic activity. On the other hand, several current and now widely received studies after evaluating the available data come to the conclusion that the so-called hiatus did not exist and that the trend line of global warming continued without weakening in the period in question.

See also

literature

  • Andrew Gettelman and Richard B. Rood: Demystifying Climate Models: A Users Guide to Earth System Models (=  Earth Systems Data and Models . Volume 2 ). Springer, Berlin Heidelberg 2016, ISBN 978-3-662-48959-8 , doi : 10.1007 / 978-3-662-48959-8 ( PDF, 6.5 MB - Open Access, is aimed at users and others who want to get an initial overview of climate models requires little specialist knowledge).
  • Hugues Goosse, PY Barriat, W. Lefebvre, MF Loutre, V. Zunz: Introduction to climate dynamics and climate modeling . 2010 ( climate.be - Open Access, early version of the textbook “Climate System Dynamics and Modeling” published by Cambridge University Press ).
  • Hans von Storch , Stefan Güss, Martin Heimann, The climate system and its modeling , Springer Verlag, Berlin (1999), ISBN 978-3540658306

Web links

Individual evidence

  1. ^ AF Prein, W. Langhans, G. Fosser, A. Ferrone, N. Ban, K. Goergen, M. Keller, M. Tölle, O. Gutjahr, F. Feser, E. Brisson, S. Kollet, J. Schmidli, NP van Lipzig, R. Leung: A review on regional convection-permitting climate modeling: Demonstrations, prospects, and challenges . In: Rev Geophys . tape 53 , no. 2 , June 2015, p. 323–361 , doi : 10.1002 / 2014RG000475 , PMID 27478878 , PMC 4949718 (free full text).
  2. ^ GA Schmidt, D. Bader, LJ Donner, GS Elsaesser, JC Golaz, C. Hannay, A. Molod, R. Neale, S. Saha: Practice and philosophy of climate model tuning across six US modeling centers . In: Geosci Model Dev . tape 10 , no. 9 , 2017, p. 3207-3223 , doi : 10.5194 / gmd-10-3207-2017 , PMID 30595813 , PMC 6309528 (free full text).
  3. Frederic Hourdin: The Art and Science of Climate Model Tuning . In: Bulletin of the American Meteorological Society . tape 98 , no. 3 , March 2017, p. 589-602 , doi : 10.1175 / BAMS-D-15-00135.1 .
  4. ^ G. Krinner, MG Flanner: Striking stationarity of large-scale climate model bias patterns under strong climate change . In: Proc. Natl. Acad. Sci. USA band 115 , no. 38 , September 2018, p. 9462-9466 , doi : 10.1073 / pnas.1807912115 , PMID 30181268 , PMC 6156650 (free full text).
  5. External Control of 20th Century Temperature by Natural and Anthropogenic Forcings; PA Stott et al., Science Vol. 290, no. 5499, pp. 2133 - 213715, December 2000 doi : 10.1126 / science.290.5499.2133 (Online)
  6. Mehl et al: Combinations of Natural and Anthropogenic Forcings in Twentieth-Century Climate; In: Journal of Climate , Vol. 17 2004 Online (PDF; 368 kB)
  7. PIK Potsdam: Ice Age Test Confirms Concerns about Future Global Warming
  8. Results from ClimatePrediction.net (PDF; 738 kB)
  9. Frank Kaspar and Ulrich Cubasch: The climate at the end of a warm period, Institute for Meteorology of the FU Berlin Online (PDF; 718 kB)
  10. S. Rahmstorf, HJ Schellnhuber: Der Klimawandel. CH Beck, 6th edition 2007, p. 42ff
  11. a b Hubertus Fischer u. a .: Palaeoclimate constraints on the impact of 2 ° C anthropogenic warming and beyond . Review Article. In: Nature Geoscience . tape 11 , 25 June 2018, doi : 10.1038 / s41561-018-0146-0 .
  12. Gerta Keller, Paula Mateo, Jahnavi Punekar, Hassan Khozyem, Brian Gertsch, Jorge Spangenberg, Andre Mbabi Bitchong, Thierry Adatte: Environmental changes during the Cretaceous-Paleogene mass extinction and Paleocene-Eocene Thermal Maximum: Implications for the Anthropocene . (PDF) In: Gondwana Research . 56, April 2018, pp. 69-89. doi : 10.1016 / j.gr.2017.12.002 .
  13. ^ Syukuro Manabe, Richard T. Wetherald (1967): Thermal Equilibrium of the Atmosphere with a Given Distribution of Relative Humidity. In: J. Atmospheric Sciences 24, pp. 241-259, quoted from: National Science Digital Library Essay about this article  ( page no longer available , search in web archivesInfo: The link was automatically marked as defective. Please check the link according to the instructions and then remove this notice.@1@ 2Template: Dead Link / nsdl.org  
  14. GCM Model information, IPCC website ( IPCC DDC: HadCM3 GCM Information )
  15. GCM Model information, IPCC website ( IPCC 4th Assessment Report (2007): Model UKMO-HADCM3 )
  16. Vicky Pope: The Hadley Center climate model HadGEM1. ( PDF ( Memento of September 17, 2007 in the Internet Archive )).
  17. ^ C. Schölzel, A. Hense: Probabilistic assessment of regional climate change in Southwest Germany by ensemble dressing , Climate Dynamics 36 (9), 2011, 2003–2014, doi : 10.1007 / s00382-010-0815-1
  18. KomPass - Introduction to climate projections / climate projections of the Federal Environment Agency . Last accessed on August 5, 2014
  19. official website of the research field interactions between geo and biosphere at the Marum
  20. Ole Bøssing Christensen, Martin Drews, Jens Hesselbjerg Christensen, Klaus Dethloff, Klaus Ketelsen, Ines Hebestadt, Anette Rinke: The HIRHAM Regional Climate Model Version 5 (beta). In: Technical Report 06-17, dmi.dk. Retrieved December 8, 2019 . Pp. 5-6.
  21. Richard B. Alley: The unstable climate, p. 12 in: Spectrum of Science Dossier 2/2005 The earth in the greenhouse
  22. Volker Mrasek: Rapid warming: Arctic climate could have irrevocably tipped , In: Spiegel-Online (Wissenschaft), accessed on December 23, 2019.
  23. Zhang, X., A. Sorteberg, J. Zhang, R. Gerdes, and JC Comiso (2008), Recent radical shifts of atmospheric circulations and rapid changes in Arctic climate system, Geophys. Res. Lett., 35, L22701, doi : 10.1029 / 2008GL035607
  24. Stroeve, J., MM Holland, W. Meier, T. Scambos, and M. Serreze (2007), Arctic sea ice decline: Faster than forecast, Geophysical Research Letters , 34, L09501, doi : 10.1029 / 2007GL029703
  25. Jochem Marotzke , Piers M. Forster: Forcing, feedback and internal variability in global temperature trends . In: Nature . tape 517 , 2015, p. 565-570 , doi : 10.1038 / nature14117 .
  26. Thomas R. Karl, Anthony Arguez, Boyin Huang, Jay H. Lawrimore, James R. McMahon, Matthew J. Menne, Thomas C. Peterson, Russell S. Vose, Huai-Min Zhang: Possible artifacts of data biases in the recent global surface warming hiatus . (PDF) In: Science . 348, No. 6242, June 25, 2015, pp. 1469–1472. doi : 10.1126 / science.aaa5632 .
  27. ^ Niamh Cahill, Stefan Rahmstorf, Andrew C. Parnell: Change points of global temperature . In: Environmental Research Letters . 10, No. 8, August 2015. doi : 10.1088 / 1748-9326 / 10/8/084002 .