Low (web) Quality - BALTEX
Low (web) Quality - BALTEX
Low (web) Quality - BALTEX
Create successful ePaper yourself
Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.
125<br />
4. Preliminary results for downscaling one CFS<br />
ensemble member with WRF<br />
Thus far we have executed two tests dynamically<br />
downscaling one CFS ensemble member for the 1993 warm<br />
season. For the first test, only standard lateral boundary<br />
nudging is applied to the outer five grid points of the model<br />
domain, equivalent to what is traditionally done in WRF<br />
numerical weather forecast applications and oftentimes in<br />
regional climate modeling applications. For the second test,<br />
both standard lateral boundary nudging and internal nudging<br />
(at all wavelengths) is applied according to the standard<br />
WRF FDDA options.<br />
Precipitation results for June 1993 are shown for the<br />
original CFS ensemble member (Figure 1), and the WRF<br />
test downscaling experiments described above (Figure 2).<br />
The precipitation for the original CFS ensemble member<br />
demonstrates that the CFS model is able to capture warm<br />
season teleconnections that lead to increased rainfall in the<br />
central U.S., consistent with the precipitation observations at<br />
this time. The WRF experiment with only lateral boundary<br />
nudging, however, produces a diminished amount of<br />
precipitation in the central U.S. as compared to the original<br />
CFS ensemble member. Thus, this experiment appears to<br />
actually take away value from CFS model. An analysis<br />
showed that the large-scale circulation patterns in this WRF<br />
experiment are significantly different than in the driving<br />
CFS model ensemble member data, in a manner consistent<br />
with Castro et al. (2005). By an improved representation of<br />
the large-scale circulation, the WRF experiment with lateral<br />
boundary nudging and internal nudging dramatically<br />
improves the forecast June precipitation. The positioning of<br />
the precipitation maximum in the central U.S. shifts slightly<br />
southward and the amount of precipitation there is increased<br />
compared to the original CFS ensemble member, more<br />
closely matching the observed precipitation. Later months<br />
in the summer (not shown) also show a much improved<br />
representation of the North American Monsoon in the<br />
Southwest U.S.<br />
System (RAMS), J. Geophys. Res., 110,<br />
D05108, doi:10.1029/2004JD004721, 2005.<br />
Castro, C.L., R.A. Pielke, Sr., J.O. Adegoke, S.D<br />
Schubert, and P.J. Pegion. Investigation of the<br />
Summer Climate of the Contiguous U.S. and<br />
Mexico Using the Regional Atmospheric<br />
Modeling System (RAMS). Part II: Model<br />
Climate Variability. J. Climate, 20, 3888-3901,<br />
2007.<br />
Miguez-Macho, G., G.L. Stenchikov, and A. Robock.<br />
Regional Climate Simulations over North<br />
America: Interactions of Local Processes with<br />
Improved Large-Scale Flow. J. Climate, 18,<br />
1227-1246, 2005.<br />
Figure 1. June 1993 CFS precipitation (mm day -1 ) from<br />
sample ensemble member.<br />
5. Conclusions and ongoing work<br />
The preliminary results of dynamically downscaling a<br />
CFS ensemble member with WRF for the warm season in<br />
North America are quite promising. Provided that the<br />
regional model is able to retain the variability in the large<br />
scale circulation fields, WRF used as a RCM can potentially<br />
add value to representation of the warm season climate.<br />
This is primarily realized by an improved representation of<br />
warm season convective precipitation. These results appears<br />
to validate the hypothesis posed by Castro et al. (2007) that<br />
RCMs can add value to the representation of warm season<br />
climate provide the driving global model produces<br />
reasonably accurate teleconnection patterns and that these<br />
are retained in the RCM. We anticipate the results here will<br />
improve further with the incorporation of spectral nudging<br />
in the CFS-WRF simulations.<br />
6. Acknowledgements<br />
This work is supported by National Science Foundation<br />
under grant number ATM-0813656. We thank Dr. Gonzalo<br />
Miguez-Macho of the University of Santiago de Compostela<br />
for providing the spectral nudging code for WRF.<br />
Selected References<br />
Castro, C.L., R.A. Pielke, Sr., and G. Leoncini, Dynamical<br />
Downscaling: Assessment of value restored and<br />
added using the Regional Atmospheric Modeling<br />
Figure 2. June 1993 CFS-WRF downscaled precipitation<br />
(mm day -1 ) using lateral boundary nudging only (top) and<br />
interior nudging (bottom).