TEST STUDY FOR COMPARISON OF PROPOSED METHOD WITH CONVENTIONAL LEVELLING METHODS In application of conventional leveling methods a test network given in figure 1a was established. Leveling measurements were performed on 5 leveling route which changes between 2.7 and 6.1km. There are 6 Istanbul GPS Network (IGNA) stations with known ellipsoidal and orthometric heights. 1 rover, 1 local reference (34082) and 2 permanent GPS stations (ISTA and TUBI) in Istanbul were used for GPS campaigns. 34082, ISTA and TUBI considered as multi-reference so, positions of the rover on 87 test points was obtained by the adjustment of the post-processed baseline vector components from 3 different reference stations (Figure 1b).  Fig.1. Network configuration for conventional and proposed leveling methods in test area Geometric Leveling measurements In GL measurements, Wild-NAK2 precise leveling instrument and millimeter-partitioned miras were used. A total of 5 routes were formed with the 6 adjacent IGNA point, the height of 87 points on the profile have been determined with the forward and reverse GL measurements. In the study area, there was no possibility observing or obtaining the gravity values due to some restriction. But as the result of the several approximate formulas (D.Clark-B.Dubuisson, Jordan-Eggert, Naebauer) the orthometric corrections that might be made in the heights measured with GL are negligible, moreover, it is seen that the lengths of leveling routes and heights differences are not too large, so orthometric corrections have been undervalued and the heights obtained as a result of raw leveling measurements have been taken as orthometric heights. Trigonometric Leveling measurements Three different trigonometric leveling methods were considered as the unidirectional, leap-frog and simultaneous reciprocal zenith angle observations. 2 full series zenith angle observed for each method. Geodimeter-520 totalstation were used for unidirectional and leap-frog observations easily. For the simultaneously and reciprocal type of trigonometric leveling two Wild-T2 theodolites were used. Using observation signs whose zenith angle series have been developed by us and mounted to the Wild-T2 theodolites, it has been simultaneously and reciprocally measured as 2 full series.  Fig.2. Trigonometric leveling measurements Index error of each theodolite and totalstation has been determined under laboratory conditions and has been brought as correction to zenith angle measurements. Distance measurements have been made with the Geodimeter-520 totalstation and in slope form and necessary corrections have been made. As shown in figure 2, measured zenith angles Z, iA and iB heights of the instruments or targets, S slope distance measured using the totalstation, equations 1,2,3 respectively UTL,LFTL,SRTL was used for the height difference between A and B points. Height difference between two adjacent points for 5 leveling routes has been computed as the sum of the several single height differences obtained from each settlement by using equations 1,2,3 [20,21,22,25].   Similar to GL differences between forward and reverse measurements and misclosure values between known and computed heights are used for a 1 km TL line, standard deviation has been computed as ± 6mm/km from misclosure and as ± 4mm/km from difference between forward and reverse measurements for SRTL. ± 6mm/km from misclosure and as ± 6mm/km from difference between forward and reverse measurements for LFTL. ± 10mml / km from misclosure and as ± 8mm/km from difference between forward and reverse measurements for UTL. GPS leveling measurements GPS receivers used in study are Ashtech Z-surveyor dual frequency, multi channel geodetic receivers. ISTA and TUBI permanent GPS stations have Ashtech Z-XII3 receivers. Each of the 87 points has been positioned from three reference stations with 3-5 minute double sessions in different satellite geometry using the post-processing kinematic method. Elevation mask is 150, epoch interval is 5 second were selected during the observations. Standard troposphere model, use of L1L2 observation and broadcast ephemerides are sufficient for the observing ellipsoidal heights in cm level. Ashtech Solution 2.6 GPS post processing software were used for processing of the achieved GPS data. First of all, processing was performed based on 34082 reference station. Then, ISTA and TUBI permanent GPS station were included the processing as second and third reference station so achieved ellipsoidal height accuracy was improved between percent 20 to 50.  |