For specific applications such as permanent GPS network calibration and national leveling network surveying, a vertical accuracy of ∼1 mm for observing durations of a few hours to a few days at maximum in 10–100-km baselines would be required. To achieve a 1-mm accuracy in height determinations with differential-GPS measurements, path delay must be corrected with an accuracy of ∼0.3 mm. This level of accuracy is not achievable with standard GPS data analysis procedures. External correction from a water vapor remote sensing technique is therefore necessary. Microwave radiometers, which have been most extensively used for this purpose, solar spectrometers, DIAL and Raman lidars are considered in this paper. The principle and performance of these techniques is reviewed in the context of wet path delay retrieving. Namely, we evaluate the errors arising during the conversion of raw measurements to wet path delay, using retrieval coefficients or standard profiles. It is shown that changes in temperature profiles can produce errors of up to 1 cm in wet path delay with microwave radiometers. Similarly, mismodeled temperature profiles can produce errors of 2–3 mm in wet path delay with DIAL and Raman lidars. Raman lidar offers the possibility to retrieve the temperature profile from total air density. Assuming that absolute concentrations of water vapor and dry gases can be retrieved, the accuracy would be unbiased. In addition, Raman lidar would also allow for the correction of hydrostatic path delay without requiring the use of mapping functions. This might reduce the residual errors due to horizontal pressure and temperature gradients. This technique will therefore be investigated in more details in a future study.