Welcome to CGD

expand collapse

CGD Climate Highlights: Past Climate Variability

What do we know about past climate variability?

Climate of Past Centuries and Millennia

Climate varies naturally in space and time. Some variations happen by chance, due to the interactions of many components that are linked in the climate system; a good example is the recurring of El Niño events, or the year-to-year variations in the location of the jet stream over the Northern Hemisphere continents in wintertime causing regionally mild or harsh winter conditions. There is also a part of climate that happens in response to certain external influences, called forcings; they include the cooling following a sudden injection of particles and gases by a big volcanic eruption, temperature and circulation changes in response to variations in the sun's output, or on longer time scales even to the changes in the Earth's orbit around the sun. Today, we are faced with the cumulative changes in greenhouse gas concentrations in the atmosphere and emissions of aerosol particles due to human activity that impose the most significant forcing on the Earth's climate. The climate response to such human influence is what we describe as Current Climate Change, and with regard to large-scale temperature, Global Warming. Change that occurs in direct response to such forcings comes superposed on the underlying natural variations. Thus, one of the most important tasks for climate scientists is to identify and separate forced (or induced) signals from the unpredictable climate "noise". This process is called "Detection and Attribution".

There are different approaches climatologists can use to study how climate responds to external influence:

They use observations and models to study directly the physical processes that translate the forcing into the climate system.
One can also reconstruct past climates and try to identify the causes of changes that did occur over past centuries and millennia.
Or in a combination of (1) and (2), one can apply models to try to simulate past climate variations and to test if the models can reproduce the elements of variability that are forced by outside factors.

In the end, the goal is to be able to identify a self-consistent geophysical framework that can explain how the best estimates of past forcings and their associated climate response on one side are reflected in the proxy-based reconstructions of climate on the other. A link between cause and effect is getting robust once simulations closely mimic real world variations in both space and time.

Climate Reconstructions

Our instrumental records only go back a little more than 100 years. At high latitudes, and particularly in the Southern Hemisphere, records are often significantly shorter. Such a limited database can generate problems in Climate Change research because it makes it much more difficult to characterize the natural climate variability in these areas, and any trends that might be present over the recent years cannot readily be associated with climate change because they could potentially occur naturally as well. One major effort undertaken by the Paleoclimate Community is to develop climate reconstructions that help to extend the short records back over many centuries. Paleoclimatology makes use of natural archives of climate information, such as the year-to-year changes in growth rings of trees, layers in ice, laminated lake or ocean sediments, and pollen or other preserved parts of plants in peat bogs; and there are also human documents referring to climatic conditions such as diaries, harvest records or descriptions of the date of first blossom of trees (a famous example is the thousand year series of the date of first blossoming of cherry trees in the Emperor's palace in Kyoto, Japan). Climate information that can be extracted from such archives is called "Proxy Data" because they are not direct measurements but indirectly inferred climate records that are used as proxies.

Most challenging are climate reconstructions of very large scales, such as the global climate. Until the data coverage increases for the Southern Hemisphere, reconstructions focus mostly on the Northern Hemisphere, where there exists much more information. The Graph {{{ WIKIPEDIA MILL-RECONSTRUCTIONS LINK to: http://upload.wikimedia.org/wikipedia/en/b/bb/1000_Year_Temperature_Comparison.png}}} shows the numerous large scale climate reconstructions currently available. They are all based on somewhat different data, spatial coverage, or the methods used to derive the hemispheric mean temperature. The range of temperature estimates at any given time by the different reconstructions also illustrates the current uncertainty in our ability to capture the past, although the general structure is actually quite similar across these different reconstructions.

Climate Models and the Recent Past

Climate models are not only run for present conditions or for future scenarios. Equally useful are simulations that attempt to reproduce past climates. Given the best estimates of all the important forcings, climate model output can be directly compared with proxy-based reconstructions. The advantage of the modeling approach is that one can vary selected forcings by removing or adding hypothesized forcing histories, explore the effect of different magnitudes of a forcing history, or look at how consistently the same signal is produced given a particular forcing episode. This flexibility also allows for a comparison of natural and human forcings. For example, the model is well suited for studies that try to explore the role of natural forcings, which explain most of the past climate variations, but how well can they explain the warming during the 20th century?

Such a study is shown here. The graph below shows what happens if the Last Millennium is simulated with three different estimates of solar irradiance changes. All simulations show periods of cooler conditions over the past centuries, often referred to as the "Little Ice Age". There are also some relatively warm intervals, particularly in the early part of the Millennium, a period sometimes called the "Medieval Warm Period", although the climatic anomalies are by far not as well established as later cooling episodes. Additionally, the warming did not appear synchronous over most parts of the world; rather they were spread over a few centuries as often quite isolated anomalous decades. This lack of a uniform significant warming had been extrapolated from Northern European records. It is also visible in the model results for that time in the early part of the Last Millennium. In contrast, the present warming is nearly uniform over the globe.

The only difference between the three simulations is the scale factor of the solar "background" variations. The best fit with proxy records over the period 850-1870 occurs if the magnitude of solar irradiance change is kept small. Larger solar irradiance changes generate too large cooling, even though the temporal structure is the same. Interestingly, in the 20th century it does not matter if the climate is 'recovering' from a very cold or a more moderate Little Ice Age. Only the inclusion of greenhouse gas changes can force the model to reproduce the observed warming.

The combination of model and proxy studies help us understand the large scale response to changes in radiative balance and thus allow us to predict with reasonable skill what future global temperatures might look like, given any particular scenario. Currently, the combination of models and reconstructions are used to establish understanding of regional climate change, which would help the public and decision makers to identify what impact climate change will have on their immediate surroundings.