thermistor readings. Differences between IRT and surface readings have been as high as 

 1°C. 



Carl D. Miller (Extemporaneous Remarks ) 



The Infrared Laboratory of the Institute of Science and Technology of the University 

 of Michigan has attempted detailed thermal mapping of Lake Erie for the Bureau of Com- 

 mercial Fisheries. This has been done from a DC-3 aircraft and operating problems have 

 been about the same as those described by others before. 



Robert A. Ragotzkie 



The Department of Meteorology of the University of Wisconsin has carried out some 

 tests with the IRT during daylight in the summer which have shown the IRT to give higher 

 readings on the average than direct surface measurements, the greatest differences being 

 about 2.0°C. The differences were shown generally to be greatest with calm water, least 

 with rough water, and intermediate with a rippled surface. 



Arthur E. Maxwell (Sponsor's Extemporaneous Comments) 



"We are impressed with the amount of work that has gone on in IRT research and 

 survey activities, but frankly we are bothered by the lack of precision attained in the pre- 

 sent application of the technique. We are also preplexed somewhat by the preoccupation 

 with bucket-depth temperatures as interpreted from the IRT, rather than the use of direct 

 infrared measurement of the microsurface. The instruments employed for most of the sur- 

 vey work to date have severe limitations, and a very critical look should be taken at the 

 working accuracy of the instrument before surveys are commenced. Of course, an under 

 standing of all sources of variations must be greatly enlarged before improvements in the 

 resolution of the instrument itself can benefit us. In the meantime, caution must be taken 

 not to alias the data. Certainly, bringing all this out for discussion this spring is most 

 timely and useful, and this perhaps will mark the turning point in the developnaent of the 

 IRT technique. Now, after some years of unsystematic collection of data, perhaps we can 

 start on the path to more sophisticated techniques in the use of the IRT for surface tern- 

 perature survey." 



In discussion, it was pointed out that we now are able to separate the phenomenological 

 problems from the instrumentation problems and set out to take these up for solution one at 

 at time. Dr. Tully pointed out that the intended use of the data should be considered before 

 undue criticisms can be drawn of the type of work so far carried out. He mentioned that for 

 many ecological uses, for studying fish distribution, etc.. we can be much more tolerant of 

 small variations than in many purely physical oceanographic studies and many military 

 uses. For instance, in studying the influence of temperature on fish distribution we are 

 generally looking for areas of sharp transition and if there is an error of a degree m ab- 

 solute temperature this may not be of great consequence. Thus, IRT survey can be useful 

 even at present levels of understanding and of instrument accuracy, in detecting ocean 

 boundaries and measuring surface temperatures in gradient areas. 



-16- 



