On the uncertainty associated with detecting global and local mean sea level drifts on Sentinel-3A and Sentinel-3B altimetry missions

Abstract

Abstract. An instrumental drift in the point target response (PTR) parameters has been detected on the Copernicus Sentinel-3A altimetry mission. It will affect the accuracy of sea level sensing, which could result in errors in sea level change estimates of a few tenths of a millimeter per year. In order to accurately evaluate this drift, a method for detecting global and regional mean sea level relative drifts between two altimetry missions is implemented. Associated uncertainties are also accurately calculated thanks to a detailed error budget analysis. A drift on both Sentinel-3A (S3A) and Sentinel-3B (S3B) global mean sea level (GMSL) is detected with values significantly higher than expected. For S3A, the relative GMSL drift detected is 1.0 mm yr−1 with Jason-3 and 1.3 mm yr−1 with SARAL/AltiKa. For S3B, the relative GMSL drift detected is −3.4 mm yr−1 with Jason-3 and −2.2 mm yr−1 with SARAL/AltiKa. The drift detected at global level does not show detectable regional variations above the uncertainty level of the proposed method. The investigations led by the altimeter experts can now explain the origin of this drift for S3A and S3B. The ability of the implemented method to detect a sea level drift with respect to the length of the common period is also analyzed. We find that the minimum detectable sea level drift over a 5-year period is 0.3 mm yr−1 at the global scale and 1.5 mm yr−1 at the 2400 km regional scale. However, these levels of uncertainty do not meet the sea level stability requirements for climate change studies.