24 M.J. BLACKWELL 



reduction factor can also be determined. Again it must be stated that these 

 techniques are only suitable for level, bare or short grass surfaces, but these 

 are nevertheless the natural surfaces which must be used first for fundamental 

 investigations of high precision. When the theoretical, or soundly-based 

 semi-empirical, advances are well estabHshed, then the simpler techniques 

 which give satisfactory results for these surfaces can be extended to other 

 crops. With the developments described above, it seems possible to obtain 

 heat-flux measurements in soil to about ± 5-10%. This will give energy 

 contributions of similar absolute accuracy to the R terms. 



We have now come to the main point of difficulty reached in all energy 

 balance studies. From eq. 7, we have : 



XE+Q= -{R+S) (9) 



where the sign convention is that fluxes directed away from the surface 

 are considered positive. The classical approach was that of Bowen (1926) 

 who expressed eq. 9 as : 



-{R+S) = XE{i + QIXE) (10) 



where QjXE is known as the Bowen ratio B. The direct measurement of Q 

 encounters difficulties similar to that for E, and the ratio B has therefore 

 received considerable study. To gain some insight into the behaviour of 

 B, we can substitute from eqs. 5 and 6 : 



dz 



where 6, the potential temperature, is used instead of T+Fz, and x is 

 replaced by pq. 



Priestley (1959) has discussed the ratio KnjKw for varying stabiHty and 

 various heights near the ground. Although this ratio varies considerably, 

 for near neutral conditions and with approach to the surface, KnlKm and 

 hence KhjKw (see section 4) is very close to unity. Within these Umits then, 

 we have : 



XB=c,f (12) 



where the gradients have been replaced by finite differences. Moreover, for 

 land surfaces in temperature cUmates, JB is sufficiently small to enable the 

 evaporation to be detern-dned with much greater accuracy than that of B 



