 |
<< Back to WCRP CMIP3 Subprojects
Assessing Climate Change using Self-Organizing Maps
| PI: |
William J. Gutowski |
| Institution: |
Iowa State University |
| Additional Investigators: |
B. Hewitson, R. Crane |
| Abstract: |
The proposed research seeks to complement work done by many on global scale climate change detection and attribution by applying a relatively new technique, Self-Organizing Maps, that is well suited for evaluating the evolution of regional climate (e.g., Gutowski et al. 2004, J. Hydromoteor.). The specific focus will be comparative regional climate change in the U.S. and southern Africa. Self-Organizing Maps (SOMs) extract typical patterns in the evolution of weather and climate and depict the region of the pattern-space continuum through which they evolve. They are especially well suited for extracting significant spatial patterns with strong, shifting gradients, as can occur with regional climate, that truncated series of orthogonal modes may distort or miss entirely. Using long-term observed and simulated climate records for the United States and southern Africa, we will examine monthly climate evolution over the past several decades to discern changes in occurrence frequency, evolution and amplitude of temperature and precipitation patterns. Quality of simulated climates will be assessed. Also of interest will be episodes or types of behavior emerging toward the end of twentieth century that are distinctly different from earlier in the century. Climate simulation output will also be used to determine links, if any, between observed trends in temperature and precipitation patterns and future scenarios, as well as to assess contributions to observed change by different forcing factors. We will use U.S. and southern African data to contrast two differing domains (different hemispheres, differing synoptic controls, and contrasting interannual variability responses to hemispheric scale processes and teleconnections). The comparative assessments of climate change and attribution collectively will allow stronger statements about the validity of identified changes and trends.
Data required: Time series of monthly average output, preferably precipitation and near-surface temperature. Ideally, output from GCM simulations that withhold some of the 20th century forcing changes will also be available. Corresponding observed data is already in-house. |
| Publications: |
Add Publication
|
<< Back to WCRP CMIP3 Subprojects
|
|
