“How Good Are Global Climate Models?” with Michael Previdi (Mar 2009)

by admin | Jul 21, 2023 | Climate Change

Originally presented 7 Mar 2009

The Jan 2006 Earth2Class featured Dr. Beate Liepert, who described her investigations into “global dimming” (http://www.earth2class.org/k12/w5_s2006/index.php).
In 2008, Dr. Michael Previdi joined Dr. Liepert’s team as a Postdoctoral Research Scientist. Since then, he and colleagues have been investigating the validity of numerical global climate models (GCMs).

Numerical computer models are the primary tools used to forecast future changes in Earth’s climate that may occur as a result of increasing amounts of manmade greenhouse gases in the atmosphere. These global climate models are based on fundamental physical principles, and they include representations of the various components of the Earth system, including the atmosphere, oceans, land surface, and, increasingly, the biosphere, or host of living organisms on Earth. In this presentation, Dr. Previdi will discuss how GCMs are developed, tested with real world measurements (e.g., of temperature), and used to predict the future evolution of climate in response to rising greenhouse gas amounts. A simple GCM freely available online will be used to demonstrate how model climate change predictions are affected by changes in the model’s physics and by different assumptions about future human activity.

For more information, go to the “Cutting-Edge Research” section of this Workshop.

Introduction to this Workshop

How Good Are Global Climate Models? Dr. Michael Previdi

Introductory ppt for this workshop

Introductory pdf for this workshop

Cutting-Edge Research

How Good Are Climate Model Predictions?

Dr. Michael Previdi

Numerical climate models are the primary tools used to forecast future changes in Earth’s climate that may occur as a result of increasing amounts of manmade greenhouse gases in the atmosphere. The confidence that we place in these models’ climate forecasts rests largely on their ability (or lack thereof) to realistically simulate the current climate and past climate change, which we evaluate using real world measurements of temperature, rainfall and other climate variables. But this is not as simple as it might sound. Measurements are often highly uncertain, and different measurements of the same variable can lead to conflicting conclusions about how the climate is changing. Additionally, some measurements are not available for a long enough period of time to confidently assess climate model performance. An illustrative example of some of these issues pertains to the change in global rainfall during the past two decades.

During the summer of 2007, Frank Wentz and colleagues published a controversial article in the journal Science claiming that the observed increase in global rainfall during the previous twenty years (1987-2006) was 2-3 times larger than simulated by climate models. They further stated that this relatively rapid rainfall increase should continue into the future, implying significant errors in climate model predictions for the 21^stcentury. These conclusions were challenged in two papers published within the last year by Michael Previdi and Beate Liepert of the Lamont-Doherty Earth Observatory (Previdi and Liepert, 2008; Liepert and Previdi, 2009). Previdi and Liepert analyzed a different set of rainfall measurements than was used by Wentz and found a smaller increase in rainfall that is in better agreement with climate models. They also showed that trends in global rainfall are likely to vary considerably from one 20-year period to the next, indicating that the observed rainfall increase during 1987-2006 may not persist into the future. This makes it more difficult to assess the realism of model predicted rainfall trends for the 21^st century.

A realistic simulation of the current climate and past changes in climate, though an important test for a climate model, does not guarantee an accurate prediction of future climate change. Such predictions are uncertain for several reasons:

i) All model simulations begin from some initial climate state, referred to as the initial condition, and then project the evolution of the climate into the future. However, the “correct” initial condition to use is never completely known due to imperfect measurements and imperfect models. Even a small change in the initial condition can cause the climate to evolve in a different manner as a result of the chaotic nature of the atmosphere (i.e., the “butterfly effect”).

ii) Certain physical processes that affect climate occur on space and time scales that are too small to be simulated by climate models. The effects of these processes must therefore be approximated usingparameterizations which introduce additional uncertainty into climate forecasts.

iii) Climate changes in response to a radiative forcing, which is a change in the amount of radiative energy available to the Earth. Increases in atmospheric greenhouse gases, variations in energy output by the Sun, and volcanic eruptions all produce a radiative forcing that causes climate to change. We don’t know how the radiative forcing of climate change will evolve in the future since we are uncertain, for example, of the extent to which humans will continue to emit greenhouse gases into the atmosphere. As such, certain assumptions about future changes in radiative forcing must be made in climate models.

In summary, climate model predictions of future climate change are uncertain because the initial conditions, physical parameterizations, and radiative forcing in these models are also uncertain. One way to quantify this uncertainty in climate change projections is to conduct many different climate model simulations in which the initial conditions, physical parameterizations, and radiative forcing are varied over physically plausible ranges. This is in fact the goal of the climateprediction.net project (http://www.climateprediction.net/), which will be discussed in more detail at the workshop.

Classroom Resources

Teaching Resources suggested by Dr. Previdi

Climateprediction.net in schools

http://www.climateprediction.net/schools/index.php

Climate-related teaching material & sample workshops involving climateprediction.net

http://www.climateprediction.net/schools/resources.php

NASA GISS Educational Global Climate Change Model

Teaching Resources from DLESE (www.dlese.org)

Abrupt Climate Change Research at LDEO	Arctic Climate Modeling Program
Energy Balance Climate Model: Stella Mac and PC	Global Climate Change (Exploring the Environment)
Global Climate Change (UCAR)

NOAA Climate Resources

NOAA Climate Program Office	Modeling Climate
NOAA 200th Celebration Top Ten: The First Climate Model	GFDL R30 Model Projected Climate Changes

Climate and Public Policy

3rd Annual “Regional Planning Collaborative: An In-Depth Discussion of Global Climate Change Initiatives in the San Diego Region.”