EVAPORATION IN THE SURFACE ENERGY BALANCE 33 



physicist and must take place before the purely numerical processes are 

 carried out by a suitable computer. As against this, it must be admitted that 

 computing aids are needed to cope with the volume of material, that 

 machines can be unbiased and objective in the processing of the data and 

 can accept a flow of data more rapidly than the human observer is able 

 to do. 



The latter point was one of the historical reasons for the development of 

 the apparatus to be described. At the end of section 3 it was mentioned that, 

 since unaspirated fme wire thermocouples have been adopted as the 

 sensing elements for temperature and humidity, their output is going to 

 consist of rapidly fluctuating voltages, following natural eddy fluctuations 

 in the frequency range to 100 c/s. 



The obvious way to record such voltages would be to use suitably 

 damped galvanometers and photographic recorders. Anyone who has 

 attempted to analyse such records would be only too willing to take the 

 next step, which is to use galvanometer-digitisers or galvanometer- 

 amphfiers. In the original apparatus described by House, Rider and Tugwell 

 (i960), the former method was in fact used to produce pulses from a photo- 

 cell which were proportional to some simple function of temperature. The 

 pulses were apphed to a pulse amplifier and could then operate electro- 

 magnetic counters directly. The electrical contacts of the Sheppard-type 

 cup anemometers were used to energise quenched relays which operated 

 the wind counters. The original system is still used for measuring the mean 

 wet- and dry-bulb temperatures, but improved methods have since been 

 adopted for the gradients themselves. 



These are now detected and ampHfied by a highly speciahsed D.C. 

 ampHfier which detennines the pulse rate of a blocking oscillator and so 

 operates the wet- and dry-bulb temperature gradient counters. At con- 

 venient stages in these circuits, there are conventional analogue computing 

 circuits which solve the psychrometric equation and also compute E and Q, 

 using the simple aerodynamic equation. The radiation and soil heat fluxes 

 are recorded by a galvanometer- amplifier which also operates counters via 

 the pulse-rate of a blocking oscillator. The counters are photographed at 

 desired intervals. The programming of the whole apparatus is determined 

 by a seconds-contact chronometer and uniselectors. 



As previously mentioned, the sensing heads which are self-directing into 

 wind have now been duphcated and readings of these are either computed 

 once per minute to give E and Q, or integrated to give mean profiles over 

 desired periods, usually of an hour. In the present writer's opinion, caution 

 is needed in interpreting the computed values. More or less instantaneous 



