Sources of Uncertainty in Global Models
By Melanie Lenart | The University of Arizona | September 14, 2008
Climate models have vastly improved in the last three decades.1 Despite measurable improvements, some issues remain that challenge the skill of Global Climate Models (GCMs).
Some of the factors complicating model projections involve:
This section refers specifically to uncertainties involving GCMs that couple atmosphere and ocean dynamics (coupled GCMs) to project changes in temperature and precipitation. Models that attempt to consider other factors, such as the changing climate’s impact on vegetation, contain even more uncertainties that are not addressed here.
Coupled GCMs serve as the starting point for a variety of other models. They must improve further before complementary models can provide realistic projections of other factors, such as climate change impacts on global vegetation and the exchange of carbon between the air, ocean and plants.2
Figure 1. Climate model resolution varies. This figure illustrates how two model grid cell resolutions (30 x 30 km and 110 x 110 km cells) depict actual precipitation patterns in Arizona (left).
| Enlarge This Figure |
Credit: PRISM Group, Oregon Climate Service, Oregon State University
Different climate processes operate on different scales in time and space. Spatial scales are often called “resolution.” Loosely comparable to the number of pixels in a digital photo, resolution relates to the number of grid cells dividing the globe for modeling purposes (Figure 1).
These scale issues affect the processes models can capture. With their relatively coarse resolution, GCMs tend to reproduce larger scale phenomenon, which include temperature fluctuations, better than smaller-scale phenomenon, such as extreme precipitation events. For instance, while heat waves tend to affect regions across several grid cells of a climate model, the intense rainfall that leads to floods often occurs at scales smaller than a grid cell.
This is one of the reasons clouds and rainfall remain challenging to model. Researchers typically must resort to using equations to approximate the activity occurring at scales smaller than the grid cell.
In some cases, improving spatial resolution can help. For instance, the GCMs with higher resolution came closer to reproducing heavy precipitation events in the continental United States.2 Similarly, the GCMs capable of simulating El Niño events more realistically tend to have more grid cells covering tropical oceans.2
However, higher resolution does not always translate into better performance. In a test of 14 different climatic factors, Thomas Reichler and Junsu Kim found only one of the three best performing GCMs had resolutions greater than a degree latitude by a degree longitude.1
Cirrus clouds (in background) are high-altitude clouds made of ice crystals. Cirrus clouds cool the Earth by reflecting incoming radiation from the sun, but also warm the Earth by absorbing outgoing radiation. Better climate change projections must account for changes in the global presence of clouds and their net cooling and warming effects.
Credit: ©Frank Roto, istockphoto.com
Clouds are a major source of uncertainty in projections for future temperature rise in GCMs. Even differences in the particles and pollutants they contain affect resulting climate projections.
For instance, one analysis found that a doubling of carbon dioxide levels could raise temperatures by 3½ to 10 degrees Fahrenheit, with the range depending on modeled differences in the size of cloud particles.3
Clouds, and their particle sizes, make analysis and projection of climate more complex for a number of reasons:
- Daytime clouds block some of the sunlight from reaching the planet’s surface, while clouds throughout the day block some of the heat from escaping from the surface. The latter effect is most detectable at night.
- The location of the cloud in the atmosphere affects whether they are more effective at blocking sunlight or trapping heat. So does their color and the size of particles and water droplets they contain.
- Smaller droplets in clouds are less likely to rain out than bigger ones. Along with precipitation rates, this can affect temperature because clouds containing smaller particles are more likely to persist, and thus continue to block some sunlight. Droplet size is affected by pollution particles, among other things.
Pollution particles can affect both temperature and precipitation in ways that are difficult to model, increasing the uncertainty in climate projections. Also known as aerosols, particles created from combustion mask the ongoing global warming to an unknown degree because of their effect on temperature and clouds.
Particles of sulfate and other pollutants tend to decrease the size of droplets in clouds. Smaller droplets tend to remain airborne longer than bigger droplets. In a study of how pollution from forest fires affected rain clouds above the Amazon, Meinrat Andreae and his colleagues found that clouds transporting smoke particles tended to persist longer.4 However, if these clouds did reach heights that allowed them to produce rainfall, the results tended to be more extreme, involving intense thunderstorms.
Aerosols have a cooling effect on temperature, too. They promote formation of smaller water droplets, which helps clouds persist while simultaneously making them better reflectors of sunlight.5 Aerosols themselves can also reflect sunlight. As a result, the recent warming has been less severe than it would have been without the shielding presence of pollution particles.
The resulting “dimming” of the sun from the ongoing influx of pollution particles is similar to the type that occurs when volcanic aerosols reach above the atmosphere’s weather layer, where they can remain airborne for a year or more.
Assessing how sensitive climate has been to the addition of pollution particles is challenging, however.
Read about the status and progress of climate modeling in Chelsey Killebrew's Q&A feature article with the Director of the National Center for Atmospheric Research.
For many decades, aerosols were increasing with greenhouse gases because they often have a common source—mainly the burning of gas, oil, and coal.
While the amount of greenhouse gases warming the planet continues to rise, the amount of sulfate particles cooling the planet is dropping over many regions, with the exception of China and other newly industrializing countries.
Since the mid-1970s, many governments have been reducing sulfate emissions because of their negative effects on human health. They provoke and worsen heart and lung diseases, including asthma.
Aerosol pollution has a shorter-lived effect on temperature than greenhouse gases. These particles typically rain out within a week or so, whereas greenhouse gases can remain airborne for many decades.
It’s unclear just how much aerosol particles have been counteracting the effect of global warming from excess greenhouse gases. However, an analysis by Andreae and colleagues reported in Nature suggests the future warming could be worse than even high-end projections if the cooling effect of aerosols have played a strong role in dampening the observed temperature rise.5
The skill of GCMs in capturing climate variability is more difficult to assess than their ability to reproduce average conditions for a few reasons.1
For one, the models are developed using data for the present climate, which doesn’t necessarily capture the full range of variability that will be found in the future. Past climates have a greater range of variability, but reconstructions of past climate from natural archives like tree-rings and ice cores contain more uncertainty than modern, instrumental obeservations of climate. This uncertaintly hinders testing of the models with past observations.
Furthermore, climate itself is complex. Ongoing research continues to shed light on the many factors influencing the variability of different climate patterns. Still, the factors themselves interact, as do climate patterns. This is particularly true for precipitation patterns affecting the Southwest.
- Reichler, T., and J. Kim. 2008. How well do coupled models simulate today’s climate? Bulletin for the American Meteorological Society, 89(3): 303-311.
- Bader, D.C., et al. 2008. Climate models: An assessment of strengths and limitations. A Report by the U.S. Climate Change Science Program and the Subcommittee on Global Change Research. Department of Energy, Office of Biological and Environmental Research, Washington, D.C.
- Senior, C.A., and J.F.B. Mitchell, 1993. Carbon dioxide and climate: The impact of cloud parameterization. Journal of Climate, 6: 5-12.
- Andreae, M.O., et al. 2004. Smoking rain clouds over the Amazon. Science, 303: 1337-1342.
- Andreae, M.O., C.D. Jones and P.M. Cox, 2005. Strong present-day aerosol cooling implies a hot future. Nature, 435: 1187-1190.