Necessity of Determining the Astronomical Refraction and Building the Refractive Delay Model Measured

According to influences of the astronomical angular refraction on classical astrometry and requirements of modern space geodesy to neutral atmospheric refractive delay corrections, the reasons why those corrections are not meet the demands of expected accuracy are analyzed, in which the main is that values of the angular refraction can not be determined directly by those methods. Aimed at the dependence of those influences on atmospheric distribution models and the requirement that the corrections should be varied with change in observing stations and observational directions, effective ways for improving accuracy of these two kinds of corrections are presented, that is, determining values of the astronomical refraction at different zenith distances in different directions at each station, building refraction models measured in different directions respectively, finding refractivity and parameters of mapping function using measured data, and constructing and adopting refractive delay corrective models varied with change in the stations and the directions. If the new technique is put in practice, in condition of not adopting atmospheric distribution model, the accuracy of the low-altitude refractive delay correction will be improved from present meter level to centimeter level, and the cutoff elevation angle will be reduced to a range of 5°. The possibility of determining values of astronomical refraction in several directions at each station is discussed.