The impact of tropospheric mapping function on PPP determination for one-month period

When using the PPP method, it is recommended to take into account the tropospheric influences for obtaining reliable estimates. Global Navigation Satellite System (GNSS) observations taken at low elevation suffer more strongly from atmospheric, antenna phase center variation and multipath effects, hence the observations are noisier than those at higher elevation angle, but they are essential to decorrelate the estimated station height and tropospheric zenith delay (ZTD). To relate the ZTD in the direction of an observation, the so-called mapping function (MF) are used. In this article the influence of different mapping function was studieds such as: Niell mapping function (NMF), Global Mapping Function (GMF) in conjunction with the Global Pressure and Temperature 2 -GPT2, Vienna Mapping Function 1 and no mapping function. The MF were used at different elevation cutoff angles – 50, 70, 100 and 150. The impact was analyzed: a) on the postfit residuals of the ionospheric free combination for phase (LC) and for pseudorange (PC), b) daily variability for North, East and Up component; c) evaluation of coordinates repeatability and how they are affected by the changes of the cutoff elevation angle and mapping function. The analyzed data was taken from 4 EUREF stations for a period of one month October 2015. By using the VMF1 mapping function, the lowest value was obtained for the postfit residuals of the LC combination for all the stations. The difference in daily variation between each individual solution for the horizontal component is at the level of ~0.3 ÷ 0.5 mm, with smaller effect on the East component compared to North, whereas the Up component is at the level of ~1.0 ÷ 1.5 mm. The standard deviation (SD) is used as a measure of station position repeatability and the results suggested that for high precision determination a cutoff elevation angle of 100 should be used. Although at low elevation 50 and 70 – the VMF1 performs better than the GMF/GPT2 and NMF, after 100 the GMF/GPT2 is strongly in agreement with VMF1 and after 150 the NMF shows similar results as VMF1 and GMF/GPT2. ARTICLE INFO