It is well-known, that GPS system upon presence data on values of such parameters as total pressure at the surface and mean temperature of troposphere allow to determine the total amount of water vapors in atmosphere. Division of total delay of GPS signals on hydrostatic and wet components is carried out upon presence data on surface pressure. Upon presence data on mean atmospheric temperature and wet delay the total amount of water vapors can be calculated. It is shown, that in the theory of atmospheric temperature measurements the presence of data on such parameters as speed of changes of vertical gradient of mean atmospheric temperature and temperature of dry surface is important. The optimization task is formulated, according which between the speeds of changes of vertical gradient of mean atmospheric temperature and temperature of dry surface should be calculated such an interrelation function upon which the integral entropy of above said parameters reaches the maximum value. The solution of the optimization task does show that the optimum function between two above said parameters has an additive shift component.