The Three-dimensional Temperature Distribution and its Variation in Time 123 

 Table 51. Temperature inversions in the western Atlantic (°C) 



t Maximum 



Also in a horizontal direction the temperature differences in the stratosphere are 

 small. The temperature distribution here must certainly be due to the stratospheric 

 circulation which starts from the locations where the stratosphere extends up to the 

 surface, that is in the polar and subpolar regions where it is in direct contact with the 

 atmosphere. The water masses that sink in these places, where the major convection 

 processes (see p. 97) originate, spread out very largely in a quasi-horizontal direction 

 towards the equator to fill up the greater part of the space underneath the troposphere 

 of the tropics and subtropics, and are thereby subjected to considerable lateral mixing 

 at the same time. 



SvERDRUP (1938) has pointed out that the stratospheric temperature distribution 

 can be mainly explained on the assumption that there is extensive lateral and vertical 

 mixing of the water masses. This mixing takes place along the isopycnic surfaces that 

 rise towards the surface in the polar and subpolar parts of the oceans. Figure 55 shows 

 that the temperature distribution in a meridional cross-section through the Atlantic 

 below 1000 m can be interpreted roughly as due to the effects of this lateral and 

 vertical mixing; the theoretical isotherms calculated from the equation on p. 108 

 taking Ax : Ay as 6 x 10^ follow a similar course than the observed isotherms. The 

 temperature distribution in the Atlantic asymmetric to the equator is partly due to 

 the effects of an inflow of warm water from the Mediterranean and partly due to the 

 strong cooling effect of the Antarctic. It cannot be doubted that mixing along the 

 isopycnic surfaces in the oceanic stratosphere is of very considerable importance 

 in the distribution of the oceanographic elements. 



(c) Adiabatic Temperature Changes and Potential Temperature 



Since sea-water is compressible, although only slightly, the pressure changes 

 undergone by a small mass of water in the ocean must be accompanied by adia- 

 batic changes in temperature which can be significant for oceanographic problems. 

 Nansen (1900, 1902) first drew attention to the thermal effects of the compressibility 

 of sea- water. If a mass of water is raised from a given depth to a shallower one, will be 

 subjected to less pressure and will expand, performing work against the external 

 pressure, and the water will be cooled by a definite amount. Analogous conditions will 

 apply for a water mass which sinks ; its temperature will increase. Since the compressi- 

 bility of water is not large these temperature changes will remain only in hmits ; how- 

 ever, since the vertical temperature gradient in the deeper layers is extremely small, 

 these adiabatic ejfects must be taken into account. 



The adiabatic temperature change Si^ for a displacement from a depth /z^ to a depth 



