Thursday, 27 January 2011: 1:30 PM
608 (Washington State Convention Center)
One important aspect to communicate climate change to users is to assess the uncertainties in projections. In this paper, the WCRP CMIP3 20th Century 20C3M experiments were examined to assess the model's ability to simulate droughts and persistent wet spells over the United States. Findings were used to diagnose the uncertainties of droughts projected by the SRESA1B scenario. Among four models: GFSL CM2.0, GFDL CM2.1, MPI_Echam5 and HadCM3, only the HadCM3 model shows that the extreme precipitation (P) events are mostly likely to persist over the western interior states similar to observations. The other models show that the preferred regions for the extreme P anomalies to persist are located over the Southern Plains and the Gulf coast. The model errors in climatology and its inability to capture the impact of ENSO on P contribute to the errors in simulating persistent events.
The occurrence of drought projected by the SRESA1B experiments indicates an increase of drought events over the California and the southern Plains. Both the differences in climatology and the impact of ENSO contribute to the increase. There are uncertainties in projecting the North American monsoon rainfall over Arizona and New Mexico. The spread was large. The mean P differences between the SRESA1B and 20C3M experiments are seasonally dependent. Most models show more rainfall in winter than summer. In reality, rainfall is dominated by the summer monsoon. These errors contribute to the uncertainties of the projection.
If CMIP 5 model simulations are available, they will also be studies.
- Indicates paper has been withdrawn from meeting
- Indicates an Award Winner