While surface temperature gradients have been highlighted as drivers of low-level atmospheric circulation, the underlying physical mechanisms remain unclear. Lindzen and Nigam (1987) noted that sea level pressure (SLP) gradients are proportional to surface temperature gradients if isobaric height (the height where pressure does not vary in the horizontal plane) is constant; their own model of low-level circulation assumed that isobaric height in the tropics is around 3 km. Recently Bayr and Dommenget (2013) proposed a simple model of temperature-driven air redistribution from which they derived that the isobaric height in the tropics again varies little but occurs higher (at the height of the troposphere). Here investigations show that neither the empirical assumption of Lindzen and Nigam (1987) nor the theoretical derivations of Bayr and Dommenget (2013) are plausible. Observations show that isobaric height is too variable to determine a universal spatial or temporal relationship between local values of air temperature and SLP. Since isobaric height cannot be determined from independent considerations, the relationship between SLP and temperature is not evidence that differential heating drives low-level circulation. An alternative theory suggests SLP gradients are determined by the condensation of water vapor as moist air converges towards the equator. This theory quantifies the meridional SLP differences observed by season across the Hadley cells reasonably well. Higher temperature of surface air where SLP is low may be determined by equatorward transport and release of latent heat below the trade wind inversion layer. The relationship between atmospheric circulation and moisture dynamics merits further investigation.

31 pages, 11 figures, 2 Tables