Assessment of RCM and urban scenarios uncertainties in the climate projections for August in the 2050s in Tokyo

This study provides the first attempt in quantifying the uncertainties in future urban climate projections due to regional climate models and metropolitan-scale urban planning scenarios. Targeted for the 2050s Augusts in Tokyo, Japan, dynamical downscaling simulations are conducted using two regional climate models, the WRF and NHRCM models, both downscaled from the global climate model MIROC5 under the RCP4.5 scenario. Both regional climate models are coupled with appropriate urban canopy models to accurately evaluate the urban climate. The projected 10-year mean temperature increases for the 2050s Augusts in the central Tokyo are roughly 2.4 A degrees C and 2.2 A degrees C, for the WRF and NHRCM models, respectively, with a roughly 0.2 A degrees C difference between the two. Urban scenario experiments with the WRF model indicate that the compact city urban scenario can reduce the August mean temperature of surrounding residential areas by 0.4 A degrees C, while the dispersed city scenario can increase the temperature by 0.1 A degrees C. On the other hand, impact of urban scenarios on the temperature increase in central Tokyo is comparative or less than the surrounding areas. The impacts of urban scenario and regional climate model differences are larger in nighttime than in daytime, but are at most 0.6 A degrees C. The results indicate that the uncertainties with the regional climate models and urban scenario are significantly less than those in emission scenarios or global climate model projections.