James Hansen - Research and Publications - Climate Model Development and Projections

Climate Model Development and Projections

Vilhelm Bjerknes began the modern development of the general circulation model in the early 20th century. The progress of numerical modeling was slow due to the slow speed of early computers and the lack of adequate observations. It wasn't until the 1950s that the numerical models were getting close to being realistic. Hansen's first contribution to numerical climate models came with the 1974 publication of the GISS model. He and his colleagues claimed that the model was successful in simulating the major features of sea-level pressure and 500mb heights in the North American region.

A 1981 Science publication by Hansen and a team of scientists at Goddard concluded that carbon dioxide in the atmosphere would lead to warming sooner than previously predicted. They used a one-dimensional radiative-convective model that calculates temperature as a function of height. They reported that the results from the 1D model are similar to those of the more complex 3D models, and can simulate basic mechanisms and feedbacks. Hansen predicted that temperatures would rise out of the climate noise by the 1990s, much earlier than predicted by other researches. He also predicted that it would be difficult to convince politicians and the public to react.

By the early 1980s, the computational speed of computers, along with refinements in climate models, allowed longer experiments. The models now included physics beyond the previous equations, such as convection schemes, diurnal changes, and snow-depth calculations. The advances in computational efficiency, combined with the added physics, meant the GISS model I could be run for five years. It was shown that global climate can be simulated reasonably well with a grid-point resolution as coarse as 1000 kilometers.

The first climate prediction computed from a general circulation model that was published by Hansen was in 1988, the same year as his well-known Senate testimony. The second generation of the GISS model was used to estimate the change in mean surface temperature based on a variety of scenarios of future greenhouse gas emissions. Hansen concluded that global warming would be evident within the next few decades, and that it would result in temperatures at least as high as during the Eemian. He argued that, if the temperature rises 0.4 °C above the 1950-1980 mean for a few years, it is the "smoking gun" pointing to human-caused global warming.

In 2006, Hansen and colleagues compared the observations with the projections made by Hansen in his 1988 testimony before the United States Congress. They described the intermediate scenario as the most likely, and that real-world greenhouse gas forcing has been closest to this scenario. It contained the effects of three volcanic eruptions in the fifty-year projections, with one in the 1995, whereas the recent Mount Pinatubo eruption was in 1991. They found that the observed warming was similar to two of the three scenarios. The warming rates of the two most modest warming scenarios are nearly the same through the year 2000, and they were unable to provide a precise model assessment. They did note that the agreement between the observations and the intermediate scenario was accidental because the climate sensitivity used was higher than current estimates.

A year later, Hansen joined with Rahmstorf and colleagues comparing climate projections with observations. The comparison is done from 1990 through January 2007 against physics-based models that are independent from the observations after 1990. They show that the climate system may be responding faster than the models indicate. Rahmstorf and coauthors show concern that sea levels are rising at the high range of the IPCC projections, and that it is due to thermal expansion and not from melting of the Greenland or Antarctic ice sheets.

Following the launch of spacecraft capable of determining temperatures, Roy Spencer and John Christy published the first version of their satellite temperature measurements in 1990. Contrary to climate models and surface measurements, their results showed a cooling in the troposphere. However, in 1998, Wentz and Schabel determined that orbital decay had an effect on the derived temperatures. Hansen compared the corrected troposphere temperatures with the results of the published GISS model, and concluded that the model is in good agreement with the observations, noting that the satellite temperature data had been the last holdout of global warming denialists, and that the correction of the data would result in a change from discussing whether global warming is occurring to what is the rate of global warming, and what should be done about it.

Hansen has continued the development and diagnostics of climate models. For instance, he has helped look at the decadal trends in tropopause height, which could be a useful tool for determining the human "fingerprint" on climate. As of 12 February 2009 (2009 -02-12), the current version of the GISS model is Model E. This version has seen improvements in many areas, including upper-level winds, cloud height, and precipitation. This model still has problems with regions of marine stratocumulus clouds. A later paper showed that the model's main problems are having too weak of an ENSO-like variability, and poor sea ice modeling, resulting in too little ice in the Southern Hemisphere and too much in the Northern Hemisphere.