When we analyze a large ensemble RCP8.5 climate change experiment we find that heat waves have become more common and intense in the Great Plains during 2070-2100 compared to 1980-2010. Much of this can be attributed to the simple direct additive effect of a 5.8°C increase in Jun-Aug surface mean temperatures in that region. But there is also a non-additive effect in that daily temperature departures from the new mean during heat waves are about 0.6°C warmer in the future epoch. Here we consider two often-proposed mechanisms by which this change in the variability of surface temperature could result from indirect influences of changes in the mean state.

One mechanism involves changes in the variability of upper tropospheric planetary waves, which we are especially interested in because we have found planetary wave structures that both affect the likelihood of heat waves and have unusually high predictability on subseasonal time scales. Our analysis does show that the amplitude of planetary wave variability has been modified in the future modeled climate. And calculations with a mechanistic model show this is indeed a consequence of the change in the mean circulation. But further analysis indicates this modification of planetary wave fluctuations is probably not responsible for the increase in Great Plains heat waves. By contrast we find changes in the magnitude of surface fluxes during heat wave events could be responsible for their strengthening and these can be attributed to the decrease in soil moisture that occurs during the future period. Hence it is changes in zonally asymmetric mean land surface quantities rather than changes in upper tropospheric fluctuations brought on by changes to the mean circulation that are of primary importance in producing the enhanced variability of surface temperature in the future climate.