On the twenty-first-century wet season projections over the Southeastern United States
What physical reasons might lead to the SEUS drying in the RSM21st simulation? Why does the RSM21st differ from CCSM3-21st? We suggest that answers to these questions will help us to reconcile the uncertainty associated with regional climate projection. In response to the first question, we propose two complementary large-scale forcing mechanisms that affect the diurnal peak rainfall: (1) the atmospheric stabilization of the SEUS environment by the westward expansion of the NASH and (2) the largescale tropical-subtropical stabilization as a response to warming of the remote tropical Pacific Ocean. Figure 8a, b, and c shows the climatological summer seasonal mean sealevel pressure from the CCSM3-20th and the CCSM3-21st and their difference, respectively. CCSM3 projects a westward shift of the NASH, as apparent from the positive values of the differences in the sea-level pressure (Fig. 8c) over a large fraction of the SEUS domain of RSM. This westward shift of the NASH in the future projection entails large-scale subsidence in the lower troposphere that is not conducive to convection. Note that Fig. 8 is plotted from the CCSM3 integrations as the domain shown is outside the regional domain of RSM used in this study.
We see in Fig. 9a, b, and c that the climatological summer seasonal mean SST from CCSM3-20th and CCSM3-21st and their differences, respectively, support the second large-scale mechanism that we hypothesize is causing the relative drying in the RSM21st simulation. The CCSM3 projects a uniform warming of the SST over most of the global oceans in the late twenty-first century for the A2 emission scenario. Of particular note is the significant warming in the equatorial Pacific, reminiscent of anomalies associated with El Niño. We expect a uniform warming of the troposphere in the tropical-subtropical latitudes as the atmospheric response to this warming of the equatorial Pacific (Yualeva et al. 1994). The pressure-weighted mean tropospheric temperature from RSM20th and RSM21st and their differences are shown in Fig. 10a, b, and c, respectively. These temperatures (in Fig. 10) are calculated by taking a pressure-weighted mean of temperatures from 850 to 200 hPa. Figure 10c confirms that tropospheric temperatures in the entire SEUS domain warm in the late twenty-first century relative to the corresponding twentieth- century climatology from the RSM integration. This relative warming of the troposphere results in a stabilization of the atmosphere that is again not conducive to atmospheric convection. To further support this argument of suppression of convection in the SEUS from these largescale stabilization mechanisms, we show the climatological summer mean surface lifted index (defined in "Appendix") from RSM20th and RSM21st and their differences in Fig. 11a, b, and c, respectively. The difference in the surface lifted index (Fig. 11c) between the two centuries is positive, suggesting the relative inhibition of convection in the late twenty-first century.
|Fig. 8 (left) a The CCSM3-20th mean summer sea-level pressure. b Similar to a but for CCSM3-21st. c The difference in the two. Units are in hPa [larger image]||Fig. 9 (right) a CCSM3-20th summer mean sea surface temperature. b Similar to a, but for CCSM3-21st. c The difference in the two. Units are in Kelvins [larger image]|
|Fig. 10 (left) a The RSM20th summer mean tropospheric temperature; units are in Kelvins. b The same as a, but for RSM21st. c The difference in the two [larger image]||Fig. 11 (right) Same as Fig. 10, but for surface lifted index. Units are in Kelvins [larger image]|
As for the differences in the projected summer precipitation between CCSM3-21st and RSM21st, we believe that the significant modulation of the diurnal variability in the latter results in uniform drying across the SEUS. Dai (2006) claimed that CCSM3 showed no improvement over CCSM2 and its erroneous depiction of diurnal variation in most of the globe, including the SEUS. Relative to the *150-km grid resolution of CCSM3, the grid resolution of 10 km in the RSM simulation is able to resolve the local orographic features, the vegetation distribution, and the coastlines far more reasonably, helping to improve the fidelity of the local diurnal variations (Stefanova et al. 2012). However, the contribution of the model physics at this resolution in the RSM may be equally important for the displayed fidelity of the RSM in terms of its diurnal variations over the SEUS. Poor rendition of the diurnal variation in rainfall negatively impacts the CCSM3 projection of late twenty-first-century summer precipitation over the SEUS. CCSM3 mean summer precipitation under the A2 scenario responds to large-scale forcing and modulation of the NASH and tropospheric stabilization by warmed equatorial Pacific SST. RSM simulations of diurnal variations in SEUS, on the other hand, respond to these largescale changes, resulting in uniform reduction in rainfall at diurnal peak, which consequently reflects in the mean summer rainfall drying in the late twenty-first century relative to the corresponding mean summer rainfall of the late twentieth century.