The aim of the thesis was the calibration and application of three different devices to measure the gas flux of methane from pockmark-structures in Lake Constance. Two of the devices measured discontinuously, the „ funnel“ and the „U-Tube“, the other device, the „spinning wheel, was measuring continuously. All of the devices had to be calibrated before their deployment in Lake Constance. The calibration of the devices showed good results. The „U-Tube“ and the „funnel“ can be used for nearly every gas flux, expect very high gas fluxes. In this case the deviation in the calibration become too big. In contrast the „spinning wheel“ can be used at lower gas fluxes. Additionally, the way of data interpretation had to be established for all devices. For the „funnel“ and the „ spinning wheel“ Microsoft Excel was chosen as a common platform. Here the filtering of the raw data for the relevant data was the most important and time-consuming part. When using the „U-tube“ with its commercial data logger the interpretation of the data was much easier und faster. This logger detected only the relevant data. But it had the disadvantage that a special device and software were necessary. The deployment of the three devices in Lake Constance revealed a different sensitivity against sediment pollution. This sedimentation had a big influence on the long-term stability of the three devices. The „U-Tube“ showed the best long-term stability of all the three devices, as shown by re-calibration in a pool after the deployment. It displayed a kind of „Self-cleaning mechanism“ during the measurement. The natural gas flux at the site of deployment for the “spinning wheel” was too high, as much gas passed the device without being detected. Only the „U-Tube“ measured realistic gas fluxes. Because of a technical defect, no data were available from the “funnel”. Overall the calculated gas flux showed a good correlation to the water level above the sensor. A higher water level and a higher water pressure caused a decrease of the gas flux. The results were observed over the long-term measurement of two months. By looking at the daily gas flux small fluctuations were observed, which were not related to the water level.