Cloud Feedbacks



Other MIPs



Google Calendar

Lab Calendar

Site Map


Privacy & Legal Notice

Thanks to Our Sponsors:

PCMDI > WCRP CMIP3 Model Output > Diagnostic Subprojects Printer Friendly Version
<< Back to WCRP CMIP3 Subprojects

A Comparative Study of Precipitation in Global Circulation Models

PI: Jennifer Alltop
Institution: Columbia University
Hurricane Katrina highlighted a lack of long-term flood planning in the United States. The disaster in New Orleans was foreseeable because of the likelihood of hurricanes in the Gulf of Mexico and the inability of the city’s levees to handle such an event. According to the 2001 Intergovernmental Panel on Climate Control (IPCC) report, as global warming continues we will likely see an increase in flood magnitudes and frequency in most regions (IPCC WG2). The report also states that in many regions, low river flows are likely to decrease leading to an increase in droughts as well. Although there is agreement in the scientific community that there will be an increase in flooding and droughts, there is very little consensus as to how much precipitation will increase or where floods and droughts will occur. Floods due to hurricanes such as Katrina are expected in low lying coastal areas prone to hurricanes but the locations of floods due to extreme rainfall events are not as easy to predict.

In an attempt to model and predict such phenomena, scientists have employed a variety of Global Circulation Models (GCMs). However, the models are often not in agreement on the places that will be affected most by an increase or decrease in precipitation. The only global visualization of annual mean precipitation change put forth in the 2001 IPCC working group 1 report was a multi-model average, which could actually obscure the individual predictions rather than pointing out their differences. To illustrate the disagreement between models, examine the figure above. Four different models were used to assess how soil moisture would change according to IPCC Scenario A2. Scenario A2 assumes global CO2 levels will approach 1000 ppm by the year 2100. The figure shows the change in summer soil moisture between the last half of the 21st century, compared to the last half of the 20th century. In some places the results are similar; in others they differ among the models. The "q-flux" models have specified ocean heat transports whereas the “dynamic” model allows ocean dynamical transports to change. Note that the scales change; the models on the left (ranges from -50 cm to +50 cm on left and -5 cm to +5 cm on right) include soil moisture below rooting depth. Therefore they have greater absolute values of soil moisture. As we can't see down to those levels, we don't know how wrong or right they are. Understanding the reasons behind the similarities and differences among these models, as well as others else where, is essential for better prediction.

I am interested in exploring how and why current GCMs differ in their precipitation predictions. I will attempt to understand the differences between models by analyzing the hydrologic changes -- precipitation, evapotranspiration, soil moisture and runoff -- from the two GISS simulations used in the upcoming 2007 IPCC report, as well as other in-house GISS models of varying climate sensitivity and physics. With this data, I will then alter the GISS GCM in a controlled manner, and see if the pattern of hydrologic change can be predicted. From there, I will analyze similar output from the other GCMs in the IPCC database to identify tendencies induced by the model physics, or by biases in various models including GISS. The biases will be determined by comparison with hydrologic (and radiative) observations of the current day climate, including output from the Tropical Rainfall Measuring Mission (TRMM) and Clouds and the Earth’s Radiant Energy System (CERES). To the extent that the forecast hydrologic change depends upon the physics or biases in each model, it will allow us to better constrain predictions of future hydrologic changes. As a graduate student at Columbia University, I have access to the NASA/GISS (Goddard Institute for Space Studies) GCM and many of its contributors, making me uniquely equipped to carry out this research.

As policy-makers continue to consult scientists on issues of global warming, how will the scientific community respond? If areas at high risk for flooding and drought could be identified, scientists, water use managers, and policy makers could work together to protect endangered areas or populations, averting future catastrophes. Even if climatologists do not always come to the same conclusions on future climate, there is a rising need for comparative analysis of GCMs to understand why scientists disagree.

Add Publication

<< Back to WCRP CMIP3 Subprojects
For questions or comments regarding this website, please contact the Webmaster.
Lawrence Livermore National Laboratory  |  Physical & Life Sciences Directorate