Assessment of height variations by GPS at Mediterranean and Black Sea coast tide gauges from the SELF projects

Abstract In the framework of the European Union SEa Level Fluctuations (SELF) I and the SELF II Projects designed to study sea level variations around the Mediterranean and Black Seas, the Global Positioning System (GPS) technique was adopted to measure the ongoing crustal movements at tide gauge stations. Tide gauges measure sea-level variations with respect to a ground benchmark. In order to determine true sea-level variations of a few millimeters per year, it is necessary to estimate the ground vertical movement to a high degree of accuracy. Countries involved in the projects were Spain, France, Italy, Greece, Bulgaria and Russia. In the period from 1993 to 1998 repeated observations at more than 28 tide gauges and at a number of reference and intermediate stations were conducted. GPS measurements were complemented by Water Vapor Radiometers (WVR) at selected sites. Individual campaigns were evaluated and analyzed and showed height variations in the millimeter–centimeter range. All data were processed following common standards using the Bernese Software in a unified analysis to generate a combined solution. Based on the normal equations of all of the campaigns, a combined solution for the SELF Projects sites was generated. In the analysis of multi-technique anchor sites to the International Terrestrial Reference Frame ITRF97, it is shown that vertical rates of different techniques may be of different sign with respect to each other and also with respect to the ITRF97 combination. Vertical rates are not reliable in general and even the restriction to the use of long the long-time series sites KOSG, ONSA, WTZR, MADR, GRAZ and MATE did not ensure significant and unique vertical rates. Nevertheless, they had to be used for the reference frame definition. The results of repeated gravity observations and the continuous GPS observations at Porto Corsini and Medicina were compared to the epoch-wise GPS campaigns and revealed that the vertical rates are in conformance with each other if taken over the same period in time. The impact of the WVR observations on tide gauge position variation estimation is shown to improve the vertical component in the case of baseline-wise observations with two WVRs by up to 50%. It is shown, that the time span is too short for epoch-type observations to resolve significant height variations. The estimated rates of about 0–20 mm/year are most likely explained by systematic and random errors in the GPS observations. The absolute gravity observations at selected tide gauges are accurate to ±3 μGal and are in general agreement with the zero result from GPS. The main outcome of the SELF campaign observations therefore was the determination of a homogeneous zero-epoch data set and the creation of a database that includes the complete link from the GPS sites to the sea level in the five Mediterranean countries involved.