UC San Diego

A series of model simulations and statistical analyses are used to examine the drivers of summer surface air temperature (SAT) variability over the continental United States (US) and the drivers of extreme heat events over Europe. An overarching goal is to improve understanding of the land-surface feedback on summer SAT, a topic of considerable socioeconomic importance. North American summer SAT variability is well-observed and known to be influenced strongly by the large-scale atmospheric circulation as well as by the land surface. Current global climate models forced with observed sea surface temperature tend to overestimate year-to-year summer SAT variations, notably in the central US, which has been identified as a "hot spot" of land-atmosphere interaction. Chapter 2 explores the hypothesis that models are overly sensitive to variations in soil moisture in the central US. Evidence is presented that links central US SAT variability in models to sensible heat flux variability, an indicator of soil moisture influence. However, the "true" influence of the land surface on SAT is not well constrained by observations and is challenging to characterize in the presence of internal atmospheric variability.

A technique called dynamical adjustment is used to separate the influence of atmospheric circulation on SAT from the influence of the land surface on SAT. It is shown in Chapter 3 that removing the effect of circulation on SAT strengthens the correlation between preseason soil moisture and SAT in the central US. Uncertainty associated with dynamical adjustment is assessed in Chapter 4, and it is confirmed that the influence of the land surface on SAT can be characterized using solely SAT and atmospheric pressure, which are well-observed fields. Extreme summer SAT in Europe is also influenced by atmospheric circulation and the land surface; the impact of preseason soil moisture on a seasonally persistent European heat wave event is assessed in Chapter 5. By superposing the same heat wave circulation pattern on an initial condition ensemble, it is shown that a heat wave following a dry spring can be up to 3˚C hotter than a heat wave following a wet spring.