The research of strategy to determine the GPS elipsoidal heights on baselines up to 10 km with Bernese

In processing the GPS data, the use of commercial processing softwares are preferred in local applications, since they give us enough accuracy. However, in the hands of experienced users with necessary changes in procedures the academic processing softwares can possibly be used in local applications just like they were used in regional and global applications. According to this approach, it is aimed to get the accurate estimation of the ellipsoidal height differences by forming short baselines up to 10km using an academic processing software. In our research project, Bernese academic GPS Software Version 4.2 which is developed by Astronomical Institute University of Berne is used to process the GPS data, and the stations that form the baselines are chosen from Southern California Integrated GPS Network on which we can observe the seismic activities. The short baselines HOLP-PMHS, BGIS-HOLP and BGIS- ELSC are formed successively in the length of about 3, 5 and 8km which are designed not to be more than 10km. They are chosen from the stations BGIS, HOLP, ELSC and PMHS. The average arithmetic values and standart deviations of the elipsoidal height differences are computed to analyze the processing results. In order to accept Bernese results considerable, they are compared with the existing results of the commercial Ashtech WinPRISM results. As for the determination of the processing procedure with Bernese GPS Software, the given example for local applications in the software Manual is accepted as a basis. By the assistance of this example but unlikely the given example an alternative processing procedure is explored by using the LI carrier instead of L3 ionosphere free linear combination and by forming a Local Ionosphere Model expecting it works better for the local networks. With these two different procedures, firstly the data intervals of 24hour long sessions, and than the data intervals of 1,2,3 and 4hour short sessions are processed with Bernese Software and the results are compared with Ashtech WinPRISM Software. As a result of this comparison, it can be obviously seen that, the developed procedure increases the quality of standart deviations but doesn't give the acceptable results in the range of 1cm which is ITRF accuracy for processing of 24hour long session data. Because of this reason another change is applied in the developed procedure. Unlikely the given example in Bernese software, but just like the procedure in Ashtech Winprism Software, the estimation of one troposphere parameter is considered instead of many troposphere parameter estimation since we assume that the estimation of more than one troposphere parameters affect the processing of 24hour long session data negatively. As a result of this, we obtained the acceptible comparisons for both Bernese and WinPRISM results. And finally; the expected accuracy in the ellipsoidal height component was obtained with academic Bernese GPS Software, when necessary changes were made in the processing procedure. Keywords: Academic and Commercial softwares, short baselines, ellipsoidal height difference, local ionosphere model, troposphere parameter. xm